Interventions for acute non‐arteritic central retinal artery occlusion

Abstract

Background

Acute non‐arteritic central retinal artery occlusion (CRAO) occurs as a sudden interruption of the blood supply to the retina and typically results in severe loss of vision in the affected eye. Although many therapeutic interventions have been proposed, there is no generally agreed upon treatment regimen.

Objectives

To assess the effects of treatments for acute non‐arteritic CRAO.

Search methods

We searched the Cochrane Central Register of Controlled Trials (CENTRAL) (which contains the Cochrane Eyes and Vision Trials Register) (2022, Issue 2); Ovid MEDLINE; Embase.com; PubMed; Latin American and Caribbean Health Sciences Literature Database (LILACS); ClinicalTrials.gov; and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP). We did not use any date or language restrictions in the electronic search for trials. We last searched the electronic databases on 15 February 2022.

Selection criteria

We included randomized controlled trials (RCTs) comparing any interventions with another treatment in participants with acute non‐arteritic CRAO in one or both eyes. 

Data collection and analysis

We used standard Cochrane methodology and graded the certainty of the body of evidence for primary (mean change in best‐corrected visual acuity [BCVA]) and secondary (quality of life and adverse events) outcomes using the GRADE classification.

Main results

We included six RCTs with 223 total participants with acute non‐arteritic CRAO; the studies ranged in size from 10 to 84 participants. The included studies varied geographically: one in Australia, one in Austria and Germany, two in China, one in Germany, and one in Italy.

We were unable to conduct any meta‐analyses due to study heterogeneity. None of the included studies compared the same pair of interventions: 1) tissue plasminogen activator (t‐PA) versus intravenous saline; 2) t‐PA versus isovolemic hemodilution, eyeball massage, intraocular pressure reduction, and anticoagulation; 3) nitroglycerin, methazolamide, mecobalamin tablets, vitamin B₁ and B₁₂ injections, puerarin and compound anisodine (also known as 654‐2) along with oxygen inhalation, eyeball massage, tube expansion, and anticoagulation compared with and without intravenous recombinant tissue plasminogen activator (rt‐PA); 4) transcorneal electrical stimulation (TES) with 0 mA versus with 66% of the participant’s individual electrical phosphene threshold (EPT) at 20 Hz (66%) versus with 150% of the participant’s individual EPT (150%) at 20 Hz; 5) ophthalmic artery branch retrograde thrombolysis versus superselective ophthalmic artery thrombolysis; and 6) pentoxifylline versus placebo.

There was no evidence of an important difference in visual acuity between participants treated with t‐PA versus intravenous saline (mean difference [MD] at 1 month −0.15 logMAR, 95% confidence interval [CI] −0.48 to 0.18; 1 study, 16 participants; low certainty evidence); t‐PA versus isovolemic hemodilution, eyeball massage, intraocular pressure reduction, and anticoagulation (MD at 1 month −0.00 logMAR, 95% CI −0.24 to 0.23; 1 study, 82 participants; low certainty evidence); and TES with 0 mA versus TES with 66% of EPT at 20 Hz versus TES with 150% of EPT at 20 Hz. Participants treated with t‐PA experienced higher rates of serious adverse effects. The other three comparisons did not report statistically significant differences. Other studies reported no data on secondary outcomes (quality of life or adverse events). 

Authors' conclusions

The current research suggests that proposed interventions for acute non‐arteritic CRAO may not be better than observation or treatments of any kind such as eyeball massage, oxygen inhalation, tube expansion, and anticoagulation, but the evidence is uncertain. Large, well‐designed RCTs are necessary to determine the most effective treatment for acute non‐arteritic CRAO.

Plain language summary

Treatments for acute central retinal artery occlusion (blockage of the blood supply to the retina of the eye)

Key messages

There was insufficient evidence to support proposed treatments for acute non‐arteritic central retinal artery occlusion.

 What did we want to find out?

Central retinal artery occlusion occurs when the blood supply to the inner part of the retina (the light‐sensitive layer inside the eye) is suddenly blocked. If the blockage is removed in time, and the blood supply returns to the retina, full recovery is possible. However, if the blockage is prolonged, retinal cells die. Various methods have been tried to remove the blockage including massaging the eye, lowering the pressure inside the eye, and dissolving clots with drugs. The best treatment for re‐establishing the blood supply is not known, and some treatments can be associated with serious adverse effects. 

What did we do?

We searched for studies that examined treatments for acute non‐arteritic central retinal artery occlusion. We summarized the results of the studies and rated our confidence in the evidence, based on factors such as study methods and sizes.

What did we find?

It was unclear whether any treatment may improve visual acuity (how clearly your eyes can see and distinguish objects) when compared with observation. Some people treated with tissue plasminogen activator (a protein involved in the breakdown of blood clots) had serious adverse effects such as bleeding into the brain tissue.

What are the limitations of the evidence?

We have little confidence in the evidence because only small number of studies were included.

How up‐to‐date is this review?
We searched for studies published up to 15 February 2022.

Summary of findings

Summary of findings 1. Tissue plasminogen activator (t‐PA) compared with standard treatment for acute non‐arteritic central retinal artery occlusion (CRAO).

Tissue plasminogen activator (t‐PA) compared with standard treatment for acute non‐arteritic central retinal artery occlusion (CRAO)
Patient or population: people with acute non‐arteritic CRAO Settings: multicenter, tertiary care hospitals, and university hospitals Intervention: t‐PA Comparison: standard treatment or intravenous saline
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Standard treatment	t‐PA
Mean change in BCVA at 1 month (logMAR)	See comment		‐	98 (2 studies)	⊕⊕⊝⊝ Low1	Change in visual acuity at 1 month MD −0.15 logMAR (95% CI −0.48 to 0.18; 16 participants) (Chen 2011); MD −0.00 logMAR (95% CI −0.24 to 0.23; 82 participants) (EAGLE)
Quality of life measures at 1 month	No data were available for this outcome.		‐	‐	‐	‐
Adverse events at 1 month	See comment		‐	146 (3 studies)	⊕⊕⊝⊝ Low1	2 neovascularization and 3 intracranial hemorrhage in t‐PA group, 1 hemiparesis and 1 death due to endocarditis in standard treatment group
*The assumed risk is based on the estimate in the control group. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). BCVA: best‐corrected visual acuity; CI: confidence interval; CRAO: central retinal artery occlusion; logMAR: logarithm of the minimum angle of resolution; MD: mean difference; t‐PA: tissue plasminogen activator
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

¹Downgraded two levels for imprecision of results.

Background

Description of the condition

Central retinal artery occlusion (CRAO) is the sudden blockage of blood flow in the central retinal artery, which normally supplies blood to the inner retina. The resulting ischemia causes a sudden and profound loss of vision. CRAO may be further classified as arteritic (associated with vascular inflammatory syndromes such as giant cell arteritis) or non‐arteritic. In non‐arteritic CRAO, the retinal blood supply may be interrupted in several ways. The artery may be blocked by an embolus (a small piece of blood clot, cholesterol, or calcium that has entered the ocular circulation, most commonly from the carotid artery or heart). Alternatively, there may be a sudden narrowing of one of the vessels that supplies the central retinal artery or of the central retinal artery itself (i.e. hemorrhage into an atheromatous plaque in the vessel wall).

The incidence of CRAO has been reported to be 1.9 per 100,000 person‐years in the United States (Leavitt 2011), 2.3 per 100,000 person‐years in the United Kingdom (Chang 2018), and 2.5 per 100,000 person‐years in Japan (Kido 2020), and is higher in men compared to women (EAGLE). CRAO is strongly associated with atherosclerotic disease, especially carotid artery stenosis (Anderson 2002; Callizo 2015); with cardiovascular risk factors such as obesity and hypertension (Callizo 2015); and with subsequent stroke, myocardial infarction, and mortality (Lavin 2018). Although many cases of CRAO have been reported in India, Nigeria, and Tunisia, there has not been research on incidence or prevalence of CRAO in low‐ and middle‐income countries (Adenuga 2015; Ratra 2012; Shilpa 2015; Shilpa 2018; Zairi 2015).

Description of the intervention

The devastating impact of CRAO on vision has led to many attempted therapies. Moreover, there is debate about how long after the artery is blocked permanent visual loss occurs; the timing of treatment is important, as shorter occlusion time may be associated with better visual recovery as measured by Snellen visual acuity (Rumelt 1999). 

A classic study showed that the primate retina suffers no detectable retinal damage with up to 97 minutes of CRAO, after which time retinal damage steadily increases; beyond 105 minutes of ischemia, irreversible retinal damage occurs (Hayreh 1980). The degree of ischemia depends on complete versus partial blockage. A recent review suggested that retinal infarction may occur after as little as 12 to 15 minutes of complete CRAO (Tobalem 2018). Additionally, return of visual function may be delayed for days after restoration of retinal circulation, further complicating the relationship between time to treatment and visual recovery (Hayreh 2018). Moreover, ischemia‐reperfusion syndrome occurs four to seven days after CRAO in rat models of retinal ischemia, further highlighting the need for prompt treatment (Saxena 2013; Sho 2005). 

Broadly, interventions for CRAO are aimed at restoring ocular blood flow as quickly as possible before the onset of ocular necrosis. Many interventions and combinations of interventions have been attempted for CRAO. These include vasodilation, physical removal of the obstruction, intraocular pressure reduction, improving red blood cell flow dynamics, enhanced external counterpulsation (EECP), and clot destruction. Systematic reviews have found conservative approaches (i.e. sublingual isosorbide dinitrate, pentoxifylline, methylprednisolone, acetazolamide, mannitol, anterior chamber paracentesis, carbogen, hyperbaric oxygen, ocular massage, and hemodilution) to be ineffective (Fraser 2009), and the American Academy of Ophthalmology Preferred Practice Pattern on CRAO does not recommend these treatments (Flaxel 2020).

How the intervention might work

Broadly, interventions for CRAO are aimed at restoring ocular blood flow as quickly as possible before the onset of ocular necrosis. Many interventions have been attempted for CRAO with different proposed mechanisms, as follows.

Observational management: Spontaneous resolution of the occlusion (e.g. embolus moving on from the ophthalmic circulation) is reported to occur in 1% to 15% of cases (Atebara 1995; Rumelt 1999; Schmidt 1992). In a recent meta‐analysis, the visual acuity of 17.7% of participants with CRAO spontaneously improved to 20/200 or better up to 48 hours after symptom onset (Schrag 2015).

Dilation of artery: Various methods have been used to try to dilate the artery, thereby allowing increased blood flow or removal of the blockage, or both. These therapies have the disadvantage of also producing systemic vasodilation and reduced systemic blood pressure. Vasodilators described in the literature include sublingual isosorbide dinitrate, Rumelt 1999, and breathing in a fixed mixture of 95% oxygen and 5% carbon dioxide (carbogen) (Deutsch 1983; Schmetterer 1996). Vasodilators such as sublingual isosorbide dinitrate work through releasing nitric oxide, which activates the enzyme soluble guanylyl cyclase pathway to act upon the actin‐myosin components in smooth muscle to decrease intracellular calcium, resulting in smooth muscle relaxation and vasodilation (Busse 1982). Carbogen is a fixed mixture of 95% oxygen and 5% carbon dioxide. It has been proposed to increase retinal blood flow and oxygenation through preventing oxygen‐induced vasoconstriction (Deutsch 1983).

Physical removal of obstruction: The aim of this treatment is to physically dislodge the embolus so that it passes through the ophthalmic circulation; however, such treatment is only effective when the cause of the obstruction is an embolus in the central retinal artery. Methods described include the patient massaging their own eye (this usually involves gently pressing on one side of the eye and then the other through closed lids), Ffytche 1974b, or the practitioner massaging the patient’s eye with fingers or a contact lens (the latter allows the retinal circulation to be viewed at the same time) (Rumelt 1999). Ocular massage has been proposed to create large fluctuations in intraocular pressure; the resulting pulsations of the retinal artery may disintegrate a thrombus or dislodge an embolus to a peripheral vessel (Ffytche 1974b)

More recently, translumenal Nd:YAG laser embolysis (TYL) or embolectomy (TYE) to selectively fragment emboli or dislodge an intravascular embolus—thus dissolving the embolus in the peripheral circulation—has been reported in non‐controlled studies to be associated with improvement in best‐corrected visual acuity (BCVA) after 24 hours (Mason 2007; Opremcak 2008). However, these approaches have significant risks such as vitreous hemorrhage and subretinal hemorrhage and have not been rigorously tested in randomized controlled trials (RCTs). 

Increasing perfusion pressure by reducing intraocular pressure: Ocular perfusion (blood flow through the eye) increases if intraocular pressure decreases. It has therefore been suggested that a sudden decrease in intraocular pressure may prompt an increase in perfusion pressure and a consequent reduction in retinal ischemia. Methods of reducing intraocular pressure described in the literature include:

• anterior chamber paracentesis: withdrawing a small amount of aqueous fluid (fluid in the front chamber of the eye) using a small needle in the anterior chamber of the eye (Ffytche 1974b). This induces a drop in intraocular pressure and dilation of the retinal arteries due to distortion of the globe (Ffytche 1974a);

• intravenous acetazolamide (Ffytche 1974b; Rassam 1993): acetazolamide is a carbonic anhydrase inhibitor that decreases intraocular pressure, increases perfusion pressure, and increases tissue partial pressure of carbon dioxide (pCO2) to cause vascular dilation (Rassam 1993);

• intravenous mannitol (Rumelt 1999): mannitol is an osmotic diuretic that reduces intraocular pressure by drawing water out of the vitreous humor (Rumelt 1999; Weber 2018);

• topical beta blockers (Feltgen 2006; Neubauer 2000): topical beta blockers reduce intraocular pressure by reducing aqueous humor production (Feltgen 2006; Neubauer 2000);

• trabeculectomy: trabeculectomy is a surgical procedure that involves the creation of an aqueous fistula (channel) between the anterior chamber and the subconjunctival space (Harvey 2000), creating a pathway for aqueous humor to flow out of the anterior chamber and thus decreasing intraocular pressure.

Improving red blood cell hemodynamics: 

Reducing red blood cell rigidity:The less deformable red blood cells are, the harder it is for them to pass through the capillaries and the greater the tissue ischemia (Iwafune 1980). Pentoxifylline is a xanthine derivative phosphodiesterase inhibitor that reduces erythrocyte rigidity, resulting in decreased blood viscosity and increased flow velocity (Seiffge 1981); it has been used in acute ischemic stroke and multi‐infarct dementia (Frampton 1995). A Cochrane Review found insufficient evidence to support its use in acute ischemic stroke (Bath 2004). Pentoxifylline has also been used orally in people with CRAO (Iwafune 1980; Kieswetter 1983), but its efficacy has not been demonstrated. 

Increasing red blood cell flow rate:Unfavorable hemodynamics (e.g. increased blood viscosity, decreased red blood cell velocity) have been suggested to contribute to vascular occlusive disease (Barnes 1980). Isovolumic hemodilution involves injection of hydroxyethyl starch or electrolyte solution to lower hematocrit to reduce erythrocyte viscosity and increase flow velocity (Wolf 1989). Retrobulbar tolazoline, an alpha‐adrenergic antagonist that promotes vasodilation and increased erythrocyte flow velocity (Rumelt 1999), has also been used to decrease resistance and increase flow (Rumelt 1999). The efficacy of these treatments in CRAO has not been demonstrated. 

Systemic steroids: Vascular endothelial edema (swelling from excess fluid) may contribute to tissue damage following CRAO. Systemic steroids such as prednisolone may reduce endothelial edema and decrease vasospasm, increasing vessel diameter and vascular flow (Hausmann 1991). Hence, intravenous steroids have been suggested as a treatment option (Hausmann 1991); however, their efficacy is unknown. 

Enhanced external counterpulsation (EECP): EECP involves three sets of pneumatic cuffs that sequentially contract at the onset of diastole and deflate before the onset of systole, resulting in improved cardiac hemodynamics including increased coronary perfusion pressure in diastole, decreased vascular resistance, and decreased cardiac workload. EECP has been used in people with ischemic heart disease and has been reported to help relieve angina, improve exercise tolerance, and improve cardiac perfusion (Arora 1999), and has been suggested as a treatment option for improving retinal ischemia and perfusion in CRAO (Sharma 2013; Werner 2004). A prospective randomized study comparing hemodilution and EECP versus hemodilution found no difference in retinal reperfusion or visual acuity after 48 hours in people with retinal artery occlusion (Werner 2004). 

Thrombolysis (dissolving the clot): Thrombolysis or fibrinolysis is a recognized treatment for acute myocardial infarctions and some types of ischemic stroke. The principle behind this treatment in CRAO is that lysing the clot will subsequently restore the ocular circulation. The thrombolytic agent may be administered intravenously or locally (i.e. intra‐arterially). 

Intravenous thrombolysis: Agents used intravenously include urokinase, Arnold 2005, and recombinant tissue plasminogen activator (rt‐PA) (Chen 2011; Pettersson 2006), serine proteases that catalyze the conversion of inactive plasminogen to the active enzyme plasmin, the primary enzyme in thrombolysis (Wiman 1978). Intravenous thrombolysis has become increasingly accepted among clinicians as an approach to CRAO treatment (Mac Grory 2021; Youn 2018); however, there are few RCTs evaluating its use. A patient‐level meta‐analysis demonstrating that intravenous alteplase (t‐PA) within 4.5 hours of symptom onset was associated with a visual recovery rate of 37.3%, Mac Grory 2020, led to several larger RCTs in progress evaluating the utility of intravenous t‐PA for CRAO (NCT03197194; NCT04526951; REVISION). Nevertheless, the efficacy of intravenous thrombolysis in CRAO treatment has not been rigorously demonstrated. 

Local intra‐arterial fibrinolysis (LIF): LIF involves superselective administration of the thrombolytic agent directly into the ophthalmic artery, theoretically reducing the risk of intracranial and systemic hemorrhage. The EAGLE study is the only prospective RCT on efficacy of LIF for CRAO (EAGLE); this trial was stopped prematurely due to no benefit compared with conservative treatment and high rate of adverse events in the LIF group. 

Combinations of treatment: Combinations of the above treatments have also been used. Kattah and colleagues conducted a non‐randomized study using anterior chamber paracentesis and intra‐arterial rt‐PA, followed by heparin with warfarin for one month and then heparin with aspirin or clopidogrel (Kattah 2002); statistical significance was not measured, but improvements in visual acuity were reported. In another non‐randomized study, Arnold and colleagues used a combination of intraocular pressure‐lowering medications, intravenous urokinase, and aspirin versus conventional treatment in their patients, finding significant improvements in visual acuity (Arnold 2005). 

Why it is important to do this review

There are no proven treatments or standardized treatment protocols for CRAO. This is exemplified by the number of different attempted treatments (Mac Grory 2021), many of which may place the patient at risk for significant complications such as intracerebral hemorrhage (EAGLE), intraocular hemorrhage (Mason 2007; Opremcak 2008), and neurologic deficits (EAGLE). 

Objectives

To assess the effects of treatments for acute non‐arteritic central retinal artery occlusion (CRAO).

Methods

Criteria for considering studies for this review

Types of studies

We considered RCTs only.

Types of participants

We considered trials that included participants with acute non‐arteritic CRAO in one or both eyes.

Types of interventions

We considered any intervention that has been used as a treatment for acute non‐arteritic CRAO. These include:

• observation (i.e. no intervention), placebo, or sham treatment;

• dilation of artery;

• physical removal of obstruction;

• increasing perfusion pressure by reduction of intraocular pressure;

• thrombolysis (dissolving the clot);

• antiplatelet therapy;

• reducing red blood cell rigidity;

• systemic steroids;

• EECP;

• any combinations of the above.

Types of outcome measures

We included studies in which the study design, participants, interventions, and comparators met our inclusion criteria irrespective of whether our prespecified outcome data were reported. 

Primary outcomes

Critical outcome

The critical outcome for this review was mean change in BCVA at one month from baseline as measured by the included studies. If a study did not report outcomes at one month, we considered the outcomes closest to one month, or at the end of study period. We used the mean values if data on mean change were not reported.

Secondary outcomes

Important outcomes

Secondary outcomes for comparison of interventions included the following:

1. quality of life measures as recorded by the study investigators;

2. adverse events (ocular effects, systemic effects)

We reported the outcome data at one month. If a study did not report outcomes at one month, we considered the outcomes closest to one month, or at the end of study period.

Search methods for identification of studies

Electronic searches

The Cochrane Eyes and Vision Information Specialist searched the following electronic databases for clinical trials. There were no language or publication year restrictions. The electronic databases were last searched on 15 February 2022.

• Cochrane Central Register of Controlled Trials (CENTRAL; 2022, Issue 2) (which contains the Cochrane Eyes and Vision Trials Register) in the Cochrane Library (searched 15 February 2022) (Appendix 1).

• MEDLINE Ovid (1946 to 15 February 2022) (Appendix 2).

• Embase.com (1947 to 15 February 2022) (Appendix 3).

• PubMed (1946 to 15 February 2022) (Appendix 4).

• Latin American and Caribbean Health Sciences Literature Database (LILACS) (1982 to 15 February 2022) (Appendix 5).

• US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov; searched 15 February 2022) (Appendix 6).

• World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp; searched 15 February 2022) (Appendix 7).

Searching other resources

We searched the reference lists of reports from identified trials for additional trials. We did not conduct manual searches of conference proceedings or abstracts specifically for this review. 

Data collection and analysis

Selection of studies

Two review authors independently assessed the titles and abstracts resulting from the electronic searches using web‐based software (Covidence). We obtained the full reports of all relevant and possibly relevant studies for assessment according to the definitions in Criteria for considering studies for this review. Any disagreements at any stage were resolved by discussion. We attempted to contact the study investigators when the information to judge study eligibility was insufficient.

Data extraction and management

Two review authors independently extracted the data using Covidence software (Covidence). Any discrepancies were resolved by discussion. 

We extracted the following data from each included study.

• Study characteristics: title, author name, sources of support, conflicts of interest.

• Methods: study design (e.g. time period, sample size, total study duration), study methodology (e.g. eligibility criteria).

• Participant characteristics: number randomized, setting, country, participant age and gender.

• Interventions: details of the intervention in each group.

• Outcomes: primary and secondary outcomes, and adverse events reported in the included studies.

• Others: associated reports including trial register.

One review author entered the data and information into RevMan Web (RevMan Web 2022). Another review author independently verified the data entry. We attempted to contact primary investigators for missing data.

Assessment of risk of bias in included studies

Two review authors independently assessed risk of bias using the RoB 2 tool, following the guidance in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019). 

We considered the following risk of bias domains:

• bias arising from the randomization process;

• bias due to deviations from intended interventions;

• bias due to missing outcome data;

• bias in measurement of the outcome;

• bias in selection of the reported result.

We judged the risk of bias for critical outcomes (i.e. mean change in BCVA at one month) as 'low risk of bias,' 'high risk of bias,' or 'some concerns.' We followed the recommended algorithms to reach an overall risk of bias assessment for each trial. Any discrepancies between review authors were resolved by discussion. 

Measures of treatment effect

We calculated the mean difference (MD) with 95% confidence intervals (CIs) for continuous outcomes which were normally distributed. In future updates, when different but similar measurements are used (e.g. quality of life measures), we will use the standardized mean difference for continuous data.

We will use the risk ratio with 95% CIs for dichotomous data (e.g. proportion of participants who experienced any adverse events) in future updates. 

Unit of analysis issues

Trials may randomize one eye or both eyes to the intervention or comparator arm. When participants were randomly allocated to treatment and only one eye per person was included in the trial, there were no unit of analysis issues. We documented how the study eye was selected (e.g. the worse eye) if this information was available. In future updates, if participants are randomly allocated to treatment, but both eyes are included and used for analysis, we will classify the trial as 'clustered' and will adjust for within‐person correlation. If the trial is a within‐person study in which one eye is randomly allocated to intervention and the other eye receives the comparator, we will analyze as paired data. 

Dealing with missing data

We attempted to contact trial investigators to obtain additional information, individual participant data, or both. When we were not able to contact authors, or we did not receive any responses within two weeks, we used the available published data. We did not impute missing data ourselves, but we will appropriately use the imputed data if computed by the trial investigators in future updates.

Assessment of heterogeneity

We examined the overall characteristics of the studies including the type of participants and interventions to assess the extent to which the studies were similar enough to make pooling of study results meaningful.

Because no meta‐analysis was conducted, we did not examine the forest plots of study results to see how consistent the results of the studies were, given the size and direction of effects. In future updates, we will calculate I², which represents the percentage of variability caused by heterogeneity rather than sampling error (chance). We will consider I² values 0% to 40% to represent insubstantial heterogeneity, 30% to 60% may represent moderate heterogeneity, 50% to 90% may represent substantial heterogeneity, and 70% to 100% may represent considerable heterogeneity (Higgins 2019). We will also consider Chi² P value due to low power with fewer studies; hence, we would consider P < 0.1 to indicate statistical significance.

Assessment of reporting biases

We assessed risk of bias in the selection of the reported result using the RoB 2 tool to look for selective or incomplete reporting. See Assessment of risk of bias in included studies.

As we did not include 10 or more studies in meta‐analysis, we did not construct funnel plots or consider tests for asymmetry for assessment of publication bias, according to Chapter 13 of the Cochrane Handbook for Systematic Reviews of Interventions (Page 2020).

Data synthesis

We performed data analysis according to Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2019). We did not perform any meta‐analyses due to heterogeneity across studies. In future updates, we will use a fixed‐effect model, which provides a more robust estimate of effect when data are sparse. We will meta‐analyze the data using a random‐effects model if more than three studies are included in future updates.

In future updates, if there is inconsistency between individual study results such that a pooled result may not be a good summary of the individual trial results (e.g. the effects are in different directions, or I² > 50% and P < 0.1), we will not pool the data. Instead, we will present a narrative summary to describe the pattern of the individual study results (Deeks 2019). If there is statistical heterogeneity, but all the effect estimates are in the same direction such that a pooled estimate would seem to provide a good summary of the individual trial results, we may present the pooled analysis.

We will follow the guidance in Chapter 23 of the Cochrane Handbook for Systematic Reviews of Interventions if we include a study with more than two intervention groups, or non‐standard design (e.g. cross‐over design) in a meta‐analysis (Higgins 2019). 

Subgroup analysis and investigation of heterogeneity

Due to lack of information, we did not perform any subgroup analysis to compare the effect of treatment on critical outcomes in the following subgroups: age, sex, symptom duration, intraocular pressure, coronary artery disease/peripheral vascular disease, atrial fibrillation or flutter, thromboembolic disease, hypercoagulability, primary hypertension, anticoagulation use, or smoking history. 

Sensitivity analysis

We planned to perform sensitivity analysis to examine the robustness of excluding the following studies:

• studies at high risk of bias in one or more domain(s);

• studies funded by industry.

However, we did not perform these sensitivity analyses because we did not conduct any meta‐analysis.

Summary of findings and assessment of the certainty of the evidence

We prepared a Table 1 for three outcomes presented in Types of outcome measures. Two review authors independently graded the overall certainty of the evidence for each outcome as one of four levels (high, moderate, low, or very low) using the GRADE classification (www.gradeworkinggroup.org/). We downgraded the certainty of the body of evidence if we identified any of the following issues.

• High risk of bias among included studies

• Indirectness of evidence

• Unexplained heterogeneity or inconsistency of results

• Imprecision of results (i.e. wide CIs)

• High likelihood of publication bias

Results

Description of studies

Results of the search

The initial electronic searches resulted in 164 titles and abstracts. We retrieved nine papers in full for further assessment, but identified no RCTs. Some studies did use control groups (Rumelt 1999; Schmidt 1992; Weber 1998), but assignment to groups was not random.

Updated searches

Updates to the searches conducted in September 2005 resulted in a further 305 titles and abstracts. We retrieved six papers in full for further assessment, of which four were rejected after full‐text assessment (Arnold 2005; Drago 1982; Gu 1999; Kattah 2002). We included two studies in the review (Incandela 2002; Werner 2004).

An updated search in June 2007 yielded a further 203 study reports. The Trials Search Co‐ordinator scanned the search results and removed any references that were not relevant to the scope of the review. Two papers were assessed for potential inclusion but were excluded based on non‐RCT study design (Holschermann 2005; Pettersen 2005).

An updated search in September 2008 yielded a further 110 reports of studies. The abstracts were scanned, but no references were relevant to the scope of the review.

An updated database search in February 2022 yielded 2198 additional unique records. After screening 2198 titles and abstracts, we retrieved 49 full‐text reports for assessment. We included five studies (17 records), identified five ongoing trials (11 records), and excluded 21 studies (21 records). We additionally excluded Werner 2004, which was included in the previous version of this review (Fraser 2009), because it included participants with CRAO and branch retinal arterial occlusion, and the results were not reported separately for these heterogeneous populations. 

Overall, we included six studies (18 records), identified five ongoing trials (11 records), and excluded 37 studies (37 records) in this update. The study flow diagram is shown in Figure 1.

1.

Study flow diagram

One large ongoing trial comparing thrombolysis with placebo started in 2022 and plans to enroll 1400 participants (REVISION). 

Included studies

For details, see Characteristics of included studies.

We included six RCTs with 223 participants, with study sizes ranging from 10 to 84 participants. The included studies were conducted in university hospitals or tertiary care hospitals in the following countries: one study in Australia, one study in Austria and Germany, two studies in China, one study in Germany, and one study in Italy. Follow‐up periods ranged from five days, Wang 2017, to six months, Chen 2011.

Interventions and comparators varied across the included studies. Chen 2011 compared tissue plasminogen activator (t‐PA) to intravenous saline. EAGLE compared local intra‐arterial fibrinolysis (LIF) using recombinant tissue plasminogen activator (rt‐PA) to isovolemic hemodilution, eyeball massage, intraocular pressure reduction, and anticoagulation. Wu 2016 applied nitroglycerin, methazolamide, mecobalamin tablets, vitamin B₁ and B₁₂ injections, puerarin and compound anisodine (654‐2) along with oxygen inhalation, eyeball massage, and anticoagulation as a standard treatment compared with and without intravenous rt‐PA. Additionally, Incandela 2002 randomized participants with sudden vision loss associated with CRAO to pentoxifylline tablets (three 600 mg tablets daily) or placebo for four weeks. Naycheva 2013 compared transcorneal electrical stimulation (TES) 0 mA versus TES with 66% of the participant's individual electrical phosphene threshold (EPT) at 20 Hz versus TES with 150% of the participant's individual EPT at 20 Hz. Wang 2017 compared ophthalmic artery branch retrograde thrombolysis with LIF.

Excluded studies

We previously excluded 16 studies, and excluded an additional 21 studies in this update. In total, we excluded 32 studies due to study design (i.e. not an RCT), four due to type of participants (i.e. not participants with CRAO), and one due to type of intervention (i.e. Chinese medicine versus Western medicine).  

For details, see Characteristics of excluded studies.

Risk of bias in included studies

We used RoB 2 to assess the following risk of bias domains for two studies that reported our critical outcome (mean change in BCVA at one month from baseline) (Chen 2011; EAGLE).

Bias arising from the randomization process

We judged both studies as at low risk of bias for this domain. In one study, the random sequence was electronically generated by the co‐ordinating center (Chen 2011). Although some baseline characteristics were unbalanced between intervention groups, this was likely due to chance as the study had a small sample size (N = 8 each group). In another study, the randomization code was produced by a Coordinating Clinical Trials Center using a computerized random‐number generator (EAGLE). Baseline characteristics in this study were comparable between groups. Allocation before randomization was appropriately concealed in both studies. 

Bias due to deviations from intended interventions

Participants and study personnel were masked to treatment assignments in Chen 2011. A trial nurse was informed of treatment allocation only at the recruiting site, with instructions for the dose of t‐PA or placebo. We judged this study as having low risk of bias. In EAGLE, participants and study investigators were unmasked to the study medications. Recruitment was ceased after the first interim analysis based on the data for the first 70 patients (planned enrollment was 200 participants) because the data indicated similar efficacy but higher risk of adverse events in the LIF group. We judged this study as having some concerns for this domain. 

Bias due to missing outcome data

Outcome data for the primary endpoint were available for all or nearly all participants (97.6%) who underwent randomization. We judged both studies as at low risk of bias. 

Bias in measurement of the outcome

Visual acuity was measured using a Snellen chart, Chen 2011, or the Early Treatment Diabetic Retinopathy Study charts according to a standardized protocol (EAGLE). It was unclear whether outcome assessors were masked, but the outcome data were not likely to be influenced by assessors knowing which intervention was received. We judged both studies as at low risk of bias for this domain. 

Bias in selection of the reported result

The outcomes and methods used to measure and analyze the outcome that were described in the protocol, Chen 2011, or trial register, EAGLE, were consistent with the published articles. We judged both studies as having low risk of bias for this domain.

Effects of interventions

See: Table 1

Tissue plasminogen activator (t‐PA) versus treatments of any kind or placebo

Three studies compared t‐PA with standard treatment, EAGLE; Wu 2016, or placebo (Chen 2011). Participants received antiplatelet therapy 24 hours after infusion.

Critical outcome

Mean change in BCVA from baseline

Two studies reported mean change in logMAR BCVA at one month after treatment (Chen 2011; EAGLE). We did not meta‐analyze the data due to the heterogeneity in comparator. Each study suggested that t‐PA had little to no effect on change in visual acuity at one month (mean difference [MD] −0.15 logMAR, 95% confidence interval [CI] −0.48 to 0.18; 16 participants (Chen 2011); MD −0.00 logMAR, 95% CI −0.24 to 0.23; 82 participants (EAGLE); Analysis 1.1, Figure 2). Wu 2016 reported the number of participants in six ordinal categories with vision ranging from no light perception to more than 0.5 based on an international standard visual acuity chart, but it did not report mean BCVA. The authors stated that the vision after treatment (time point unspecified) was better in the treatment group compared with the control group (P < 0.05). 

1.1. Analysis.

Comparison 1: Tissue plasminogen activator (t‐PA) versus standard treatment, Outcome 1: Mean change in best‐corrected visual acuity from baseline (logMAR)

2.

We judged the certainty of the evidence as low, downgrading two levels for imprecision.

Important outcome

Quality of life measures

No studies reported this outcome.

Adverse events

One (12.5%) participant in Chen 2011 and two (4.8%) participants in EAGLE treated with t‐PA experienced intracranial hemorrhage. Chen 2011 also reported that two (13%) participants developed neovascularization of the retina. In EAGLE, one (2.5%) participant in the standard treatment group suffered from a right hemiparesis and reduced consciousness one day after treatment. One (2.5%) participant in this group died due to endocarditis, which the authors claimed was not treatment related. At least one minor adverse event was reported in one (2.1%) participant in the t‐PA group and 12 (34.3%) participants in the standard treatment group. Wu 2016 reported that three (12.5%) participants in the t‐PA group experienced bleeding gums after receiving intravenous thrombolysis. 

We judged the certainty of the evidence as low, downgrading two levels for imprecision.

Other outcomes

Wu 2016 reported that participants treated with intravenous rt‐PA thrombolysis and other therapies (nitroglycerin, methazolamide, mecobalamin tablets, vitamin B₁ and B₁₂ injections, puerarin and compound anisodine [654‐2] along with oxygen inhalation, eyeball massage, and anticoagulation) had significantly improved visual field compared with participants who received only the other standard treatment without intravenous rt‐PA thrombolysis. 

Other comparisons

None of the included studies compared the same pair of interventions, thereby precluding meta‐analysis.

Incandela 2002 showed an increase in retinal blood flow velocity and a decrease in analogue score for clinical improvement (0 to 10) in both cases and controls. The authors reported a greater increase in blood flow parameters in cases after pentoxifylline usage compared to controls and a "significant" difference in analogue score decrease with a greater decrease in the treated group. As both groups contained very small numbers, it is difficult to see how statistical significance could be reached. The paper did not provide any details of changes in visual acuity in cases or controls. The paper does not discuss any adverse events associated with the treatment. There was no discussion of the economic aspects of the trial.

Naycheva 2013 found no statistically significant differences in visual field between TES with 0 mA, TES with 66% of the participant's individual EPT at 20 Hz, and TES with 150% of the participant's individual EPT at 20 Hz. However, the a‐wave slope of the high‐intensity flash under scotopic conditions increased during follow‐up in the 150% treatment group and decreased in the 66% and sham treatment groups. No serious adverse effects were reported, although three participants reported foreign‐body sensation after TES.

Wang 2017 found that ophthalmic artery branch retrograde thrombolysis and LIF were similarly effective, as measured by postoperative digital subtraction angiography, fundus fluorescein angiography, and visual acuity tests. No serious adverse effects were reported for either group.

Discussion

Summary of main results

Based on data from six RCTs of treatments for CRAO, there was no significant difference in visual acuity between t‐PA and placebo; t‐PA and isovolemic hemodilution, eyeball massage, intraocular pressure reduction, and anticoagulation; ophthalmic artery branch retrograde thrombolysis and LIF; and TES with 0 mA, TES with 66% of EPT at 20 Hz, and TES with 150% of EPT at 20 Hz. 

We identified that TES with 0 mA and 66% of EPT at 20 Hz decreased the a‐wave slope of the high‐intensity flash under scotopic conditions and increased the slope for 150% of EPT at 20 Hz. Ophthalmic artery branch retrograde thrombolysis and LIF had statistically similar effects based on postoperative digital subtraction angiography and fundus fluorescein angiography. Participants treated with intravenous rt‐PA thrombolysis and other standard treatments significantly improved visual fields compared with those who received only the standard treatment without intravenous rt‐PA thrombolysis.

Two of three studies reported serious adverse effects, including intracranial hemorrhage, after t‐PA usage. Serious adverse events were not reported after TES or ophthalmic artery branch retrograde thrombolysis. Studies did not report tracking adverse effects for Nd:YAG laser embolysis and pentoxifylline tablets.

Overall completeness and applicability of evidence

The findings of this review are applicable to people with acute non‐arteritic CRAO and the types of interventions used in the included trials, which were conducted in Australia, Austria, China, Germany, and Italy. However, limitations include small sample sizes and study heterogeneity in the types of intervention and outcome assessment. Additionally, studies may have included people with ophthalmic artery occlusion. We attempted to contact study investigators for additional data but received no further information. 

Quality of the evidence

The overall certainty of the evidence for mean change in BCVA at one month (critical outcome) and adverse events (other important outcomes) was low due to the imprecision of results. No data were available for quality of life measures at one month.

Potential biases in the review process

We followed standard Cochrane methodology to minimize bias in the conduct of this systematic review. An experienced Information Specialist at Cochrane Eyes and Vision performed highly sensitive searches. We used Methodological Expectations of Cochrane Intervention Reviews (MECIR) standards for reporting new Cochrane Intervention Reviews (editorial-unit.cochrane.org/mecir). Two review authors, one of whom was an experienced Cochrane Eyes and Vision Information Specialist, independently performed the title and abstract screening, full‐text review, data extraction, and risk of bias assessments. None of the review authors had any financial conflicts of interest.

Agreements and disagreements with other studies or reviews

We did not identify previous systematic reviews that investigated interventions for non‐arteritic CRAO using RCTs. Available systematic reviews and meta‐analyses on interventions for CRAO were limited by small sample sizes, varied study designs, non‐standardized treatment and control interventions, and lack of specificity regarding timeframe of intervention administration and follow‐up.

Dumitrascu 2020's systematic review found improved visual outcomes in people administered intravenous and intra‐arterial thrombolysis within 4.5 hours of symptom onset; however, the included studies were limited by heterogenous reporting of visual outcomes, low percentage of patients receiving intravenous thrombolysis within 4.5 hours of symptom onset, and small sample sizes. A systematic review by Noble 2008 did not find sufficient evidence to support the routine use of intra‐arterial thrombolysis for CRAO. Schrag 2015's meta‐analysis, which included a total of 15 studies, demonstrated that fibrinolysis was beneficial within 4.5 hours or less of symptom onset; however, the study was limited by different populations of control and treatment arms and non‐standardized control interventions. Huang 2022's meta‐analysis of eight studies found that people receiving intravenous thrombolysis had better BCVA at various follow‐up times compared with those not receiving intravenous thrombolysis; however, there was no difference in final BCVA between groups. A meta‐analysis by Rosignoli 2022 on the effects of hyperbaric oxygen in people with CRAO demonstrated no effect on final follow‐up visit visual acuity. 

A meta‐analysis of five retrospective studies and one RCT by Page 2018 found a significant improvement in visual acuity after intra‐arterial thrombolysis; however, the authors did not define a timeframe by which to measure visual improvement. A meta‐analysis by Wang 2021 similarly demonstrated a significant improvement in visual acuity after intravenous thrombolysis in people with CRAO; however, the study did not define a timeframe for thrombolysis administration or a timeframe by which to measure visual improvement (e.g. participants' visual acuity may have spontaneously improved over time). 

Authors' conclusions

Implications for practice.

Based on low‐certainty evidence, we found no significant difference in visual acuity when tissue plasminogen activator (t‐PA) or transcorneal electrical stimulation (TES) was compared with placebo or existing treatments such as eyeball massage, oxygen inhalation, tube expansion, and anticoagulation. Oral pentoxifylline significantly improved retinal perfusion when compared with placebo, although the effect on visual acuity was unknown. Due to the small sizes of most included studies and low certainty of evidence, there is insufficient evidence to support the evaluated interventions for people with non‐arteritic central retinal artery occlusion (CRAO). There were higher rates of serious adverse effects for t‐PA.

For people with non‐arteritic CRAO or ophthalmic artery occlusion, a systematic etiological workup is important to evaluate for carotid occlusive and thromboembolic disease (AAO 2019). People with acute non‐arteritic CRAO or ophthalmic artery occlusion should be immediately referred to the nearest stroke center for urgent evaluation due to their increased risk for stroke (AAO 2019; Mac Grory 2021).

Implications for research.

Further research is necessary to identify effective interventions for CRAO using rigorous study design, measuring clinically significant outcomes (best‐corrected visual acuity [BCVA], quality of life, adverse events), and with larger sample sizes. Study investigators should report trials transparently using the CONSORT Statement.

What's new

Date	Event	Description
27 January 2023	New citation required but conclusions have not changed	Four studies newly included (Chen 2011; EAGLE; Mehboob 2021; Naycheva 2013; Wang 2017). Relevant sections have been updated. Werner 2004 was excluded.
27 January 2023	New search has been performed	Electric searches updated.

History

Protocol first published: Issue 1, 2000
Review first published: Issue 1, 2002

Date	Event	Description
5 November 2008	New citation required but conclusions have not changed	Issue 1 2009: Enhanced external counterpulsation (EECP) has been added as an intervention and 2 trials have been included (Incandela 2002; Werner 2004).
23 October 2008	New search has been performed	Searches updated.
10 April 2008	Amended	Converted to new review format.

Risk of bias

Risk of bias for analysis 1.1 Mean change in best‐corrected visual acuity from baseline (logMAR).

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Chen 2011

Low risk of bias

Patients were randomized 1:1 in a block design via a Website run by the coordinating center at Flinders Medical Centre. A concealed printout of the treatment allocation was passed on to a trial nurse at the recruiting site, with instructions for the dose of t‐PA or placebo that needed to be prepared. Participants who were assigned to placebo had significantly less time to presentation and treatment delivery compared with those who were assigned to the intervention (P=0.04 and P=0.01, respectively). However, this may be due to small sample size (N=8 each group).

Low risk of bias

A concealed printout of the treatment allocation was passed on to a trial nurse at the recruiting site, with instructions for the dose of t‐PA or placebo that needed to be prepared. Thus, the treating stroke team was blinded to treatment allocation. Ophthalmologist who evaluated patients on day 1 was masked to the allocation. Intention‐to‐treat analysis was followed.

Low risk of bias

There were no missing outcome data.

Low risk of bias

Visual acuity was measured by Snellen chart. Participants were examined by an ophthalmologist who were masked to treatment allocation on day 1 to document visual acuity. It was unclear if outpatient ophthalmologist and neurologist were masked at 1 month follow‐up. However, the outcome data were not likely to be influenced by knowledge of intervention received.

Low risk of bias

All prespecified outcomes in protocol were reported in the final paper. Protocol: Nonparametric techniques (Mann–Whitney) will be used to compare t‐PA and placebo groups on the primary outcome (i.e. visual acuity) Final analysis: A Mann‐Whitney U test was used for analyzing continuous ordinal, nonparametric variables owing to the nonhomogeneity of the variances between the two groups in the sample population.

Low risk of bias

See above.

EAGLE

Low risk of bias

The randomization code was produced by a Coordinating Clinical Trials Center using a computerized random‐number generator. The study center was used as a stratification criterion, and a 1:1 block randomization with randomly varying block size was undertaken within each center. Randomization took place using sealed and opaque envelopes with consecutive patient numbers to be opened in a given order indicating the therapeutic method to be used (CST, LIF). Adherence to the guidelines in handling the envelopes was verified by clinical monitors, ensuring concealment of the randomized treatment. Baseline characteristics were comparable between groups.

Some concerns

Participants and study personnel could not be masked due to the nature of interventions. The Data Monitoring Committee (DMC) recommended that recruitment cease after the first interim analysis based on the data of the first 70 patients, which indicated similar efficacy and higher rate of adverse reactions. Because patient recruitment was not suspended during conduction of the interim analysis, 14 additional subjects were recruited before recruitment was finally stopped. The full analysis set (FAS) included all randomized patients for whom a treatment or parts of a treatment had been performed, irrespective of whether the treatment had followed the protocol. Patients were analyzed as belonging to the randomized treatment arm according to intention to treat.

Low risk of bias

The primary efficacy analysis was performed with a Wilcoxon test at a 2‐sided alpha level of 0.05 in the FAS. The FAS included all randomized patients for whom a treatment or parts of a treatment had been performed, irrespective of whether the treatment had followed the protocol. Patients were analyzed as belonging to the randomized treatment arm according to intention to treat. In case of missing BCVA values at the 1‐month visit, the value was replaced using the last‐observation‐carried‐forward principle.

Low risk of bias

The primary efficacy end point was the change in BCVA 1 month after therapy compared with before therapy. This was measured by the difference from baseline logMAR at the 1‐month visit. The logMAR was assessed with the Early Treatment Diabetic Retinopathy Study charts in a standardized protocol. Although it was unclear if outcome assessors were masked, the outcome data were not likely to be influenced by knowledge of intervention received.

Low risk of bias

Primary outcome of visual acuity 1 month after therapy from baseline measured according to ETDRS Scale was specified in trial registry and reported in the published article.

Some concerns

See above.

Appendices

Appendix 1. CENTRAL search strategy

#1 MeSH descriptor Retinal Artery
#2 MeSH descriptor Retinal Artery Occlusion
#3 retina* near arter*
#4 occlus* or obstruct* or clos* or stricture* or steno* or block* or embolism*
#5 (#3 AND #4)
#6 CRAO
#7 (#1 OR #2 OR #5 OR #6)

Appendix 2. MEDLINE search strategy

1. Randomized Controlled Trial.pt.
2. Controlled Clinical Trial.pt.
3. (randomized or randomised).ab,ti.
4. placebo.ab,ti.
5. drug therapy.fs.
6. randomly.ab,ti.
7. trial.ab,ti.
8. groups.ab,ti.
9. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8
10. exp animals/ not humans.sh.
11. 9 not 10
12. exp retinal artery/
13. exp retinal artery occlusion/
14. (retina* adj3 arter*).tw.
15. (occlus* or obstruct* or clos* or stricture* or steno* or block* or embolism*).tw.
16. 14 and 15
17. CRAO.tw.
18. 12 or 13 or 16 or 17
19. 11 and 18

The search filter for trials at the beginning of the MEDLINE strategy is from the published paper by Glanville (Glanville 2006).

Appendix 3. EMBASE search strategy

#1. 'randomized controlled trial'/exp 
#2. 'randomization'/exp 
#3. 'double blind procedure'/exp 
#4. 'single blind procedure'/exp 
#5. random*:ab,ti 
#6. #1 OR #2 OR #3 OR #4 OR #5 
#7. 'animal'/exp OR 'animal experiment'/exp 
#8. 'human'/exp 
#9. #7 AND #8 
#10. #7 NOT #9 
#11. #6 NOT #10 
#12. 'clinical trial'/exp 
#13. (clin* NEAR/3 trial*):ab,ti 
#14. ((singl* OR doubl* OR trebl* OR tripl*) NEAR/3 (blind* OR mask*)):ab,ti
#15. 'placebo'/exp 
#16. placebo*:ab,ti 
#17. random*:ab,ti 
#18. 'experimental design'/exp 
#19. 'crossover procedure'/exp 
#20. 'control group'/exp 
#21. 'latin square design'/exp 
#22. #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21
#23. #22 NOT #10 
#24. #23 NOT #11 
#25. 'comparative study'/exp 
#26. 'evaluation'/exp 
#27. 'prospective study'/exp 
#28. control*:ab,ti OR prospectiv*:ab,ti OR volunteer*:ab,ti
#29. #25 OR #26 OR #27 OR #28 
#30. #29 NOT #10 
#31. #30 NOT (#11 OR #23) 
#32. #11 OR #24 OR #31 
#33. 'retina artery'/exp 
#34. 'retina artery occlusion'/exp 
#35. (retina* NEAR/3 arter*):ab,ti,kw 
#36. occlus*:ab,ti,kw OR obstruct*:ab,ti,kw OR clos*:ab,ti,kw OR stricture*:ab,ti,kw OR steno*:ab,ti,kw OR block*:ab,ti,kw OR embolism*:ab,ti,kw
#37. #35 AND #36 
#38. crao:ab,ti,kw 
#39. #33 OR #34 OR #37 OR #38 
#40. #32 AND #39 

Appendix 4. PubMed search strategy

1. ((randomized controlled trial[pt]) OR (controlled clinical trial[pt]) OR (randomised[tiab] OR randomized[tiab]) OR (placebo[tiab]) OR (drug therapy[sh]) OR (randomly[tiab]) OR (trial[tiab]) OR (groups[tiab])) NOT (animals[mh] NOT humans[mh])
2. (retina*[tw] AND arter*[tw]) 
3. occlus*[tw] OR obstruct*[tw] OR clos*[tw] OR stricture*[tw] OR steno*[tw] OR block*[tw] OR embolism*[tw]
4. #2 AND #3
5. crao[tw] 
6. #4 OR #5
7. #1 AND #6
8. Medline[sb]
9. #7 NOT #8

Appendix 5. LILACS search strategy

(MH:A07.015.114.765$ OR MH:A07.015.611.647$ OR MH:C11.768.400$ OR MH:C14.907.137.780$) OR ((retina$ AND arter$) AND (occlus$ or obstruct$ or clos$ or stricture$ or steno$ or block$ or embolism$)) OR CRAO

Appendix 6. ClinicalTrials.gov search strategy

((retina AND artery) AND (occlusion OR occlusions OR occlusive OR obstruct OR obstruction OR obstructions OR obstructive OR obstructed OR close OR closed OR closure OR stricture OR stenosis OR stenotic OR stenoses OR stenosed OR block OR blocker OR blockade OR blocked OR blocking OR embolism)) OR CRAO

Appendix 7. WHO ICTRP search strategy

retina AND artery OR retinal AND artery OR retina AND arterial OR retinal AND arterial OR CRAO

Data and analyses

Comparison 1. Tissue plasminogen activator (t‐PA) versus standard treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Mean change in best‐corrected visual acuity from baseline (logMAR)	2		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Chen 2011.

Study characteristics
Methods	Study design: parallel‐group randomized controlled trial Number randomized (total and per group): 16 participants in total; 8 participants each group Unit of randomization (individual or eye): individual  Number analyzed (total and per group): 16 participants in total; 8 participants each group Unit of analysis (individual or eye): individual Exclusions and losses to follow‐up (total and per group): 1 (discontinued due to intracerebral hemorrhage) in t‐PA group How were missing data handled?: not reported  Length of follow‐up: 6 months Reported power calculation (Y/N), if yes, sample size and power: Y, "With these figures (48.5% in treatment versus 4.8% in placebo), a sample size of 25 in each group was required to demonstrate an absolute difference of 43.7% in the proportion of subjects achieving a 3‐line improvement in Snellen VA with t‐PA compared with controls at an alpha of 0.05 and power of 80%."
Participants	Country: Australia Setting: tertiary care hospitals (2 sites) Baseline characteristics: Tissue plasminogen activator (t‐PA), n = 8 Age (mean ± SD, range): 73 ± 8 years Gender: 6 men and 2 women Time from symptoms to diagnosis/presentation: 9.1 ± 6.1 hours Time from symptoms to treatment: 14.4 ± 6.5 hours Placebo, n = 8 Age (mean ± SD, range): 67 ± 9 years Gender: 5 men and 3 women Time from symptoms to diagnosis/presentation: 3.9 ± 2.7 hours Time from symptoms to treatment: 7.3 ± 3.0 hours Overall, n = 16 Age (mean ± SD, range): 70 ± 9 years Gender: 11 men and 5 women Time from symptoms to diagnosis/presentation: not reported Time from symptoms to treatment: not reported Inclusion criteria: age ≥ 18 acute CRAO within 24 hours of onset of symptoms (i.e. within 24 hours of last known time with normal vision); a presumed thromboembolic cause; no evidence of temporal arteritis by clinical assessment or laboratory studies (e.g. erythrocyte sedimentation rate); non‐contrast CT brain demonstrating no acute intracranial hemorrhage, infarction, or mass lesion, and CT angiography demonstrating no ipsilateral carotid artery occlusion Exclusion criteria: a history of intracerebral hemorrhage at any stage; history of ischemic stroke within the last 3 months; history of systemic hemorrhage within the last 3 months; inability to obtain informed consent; pregnancy, and clinical, biochemical or imaging predictors of an increased risk of intracerebral hemorrhage including: major surgery or trauma within 2 weeks; gastrointestinal or urinary bleeding within 3 weeks; arterial puncture or lumbar puncture within 7 days; a platelet count of < 100,000/mL or who had heparin administered within the last 48 hours or vitamin K antagonist with an INR of > 1.6, systolic blood pressure of > 185 mmHg and/or a diastolic blood pressure of > 110 mmHg, or a serum glucose value > 22 mmol/L Baseline equivalence: participants who were assigned to the placebo group had a shorter mean time to presentation and treatment delivery compared with the t‐PA group (P = 0.04 and P = 0.01, respectively)
Interventions	Tissue plasminogen activator (t‐PA): "Alteplase, 0.9 mg/kg intravenous (maximum, 90 mg). A 10% bolus was given over 1 minute, followed by the remaining dose over 1 hour. Antiplatelet therapy was commenced 24 hours after infusion in both groups after a CT brain scan was obtained to exclude cerebral hemorrhage." Placebo: "10 mL of normal saline in a syringe administered over 1 minute, followed by 50 mL of normal saline given as an infusion over 1 hour. Antiplatelet therapy was commenced 24 hours after infusion in both groups after a CT brain scan was obtained to exclude cerebral hemorrhage."
Outcomes	Outcomes reported: improvement in Snellen VA by ≥ 3 lines between baseline and 6 months, equating to a > 0.3 change in the logMAR vision score; the mean improvement in VA stratified according to the following time intervals from symptom onset to treatment: 0 to 6 hours, 6 to 12 hours, and 12 to 24 hours Adverse outcomes: serious adverse events, intracranial hemorrhage Measurement time points: 1, 3, 6 months Other issues with outcome assessment (e.g. quality control for outcomes, if any): none
Notes	Study period: recruitment from 1 July 2008 to 1 April 2010 Publication language: English Trial registration: ACTRN12608000441314 Conflicts of interest: "A.W.L. was supported by a National Health and Medical Research Council, National Institutes of Clinical Studies Fellowship" Funding source: "The study received research grant support from the Perpetual Trustee The Lindsay & Heather Payne Medical Research Charitable Foundation."

EAGLE.

Study characteristics
Methods	Study design: parallel‐group randomized controlled trial Number randomized (total and per group): 84 eyes of 84 participants in total; 44 participants in LIF and 40 participants in CST Unit of randomization (individual or eye): individual  Number analyzed (total and per group): 82 eyes of 82 participants in total; 42 participants in LIF and 40 participants in CST (intention‐to‐treat analysis) Unit of analysis (individual or eye): individual (1 eye per participant was included) Exclusions and losses to follow‐up (total and per group): 2 participants in LIF (no treatment was initiated, and the reason was not stated) How were missing data handled?: last‐observation‐carried‐forward principle Length of follow‐up: 1 month Reported power calculation (Y/N), if yes, sample size and power: Y,  "One hundred patients in each group were required to detect this difference (corresponding to an odds ratio of 2.2) between treatment groups using a 2‐sided test with alpha level of 0.05 with 80% power"
Participants	Country: Austria and Germany Setting: multicenter including university hospitals (9 sites) Baseline characteristics: LIF using rt‐PA, n = 42 Age (mean ± SD, range): 63 ± 10 years Gender: 31 men and 11 women Time from symptoms to diagnosis/presentation: 9.5 ± 6.02 hours Time from symptoms to treatment: 12.78 ± 5.77 hours CST, n = 40 Age (mean ± SD, range): 63 ± 9 years Gender: 29 men and 11 women Time from symptoms to diagnosis/presentation: 9.14 ± 5.34 hours Time from symptoms to treatment: 10.99 ± 5.49 hours Overall, n = 82 Age (mean ± SD, range): 63 ± 9 years Gender: 60 men and 22 women Time from symptoms to diagnosis/presentation: 9.33 ± 5.67 hours Time from symptoms to treatment: 11.89 ± 5.67 hours Inclusion criteria: age between 18 and 75 years, non‐arteritic CRAO with symptoms for less than 20 hours, and BCVA worse than 0.5 logMAR, which is a Snellen equivalent of 20/63 (determined with the use of an Early Treatment Diabetic Retinopathy Study chart) Exclusion criteria: central retinal artery occlusion lasting longer than 20 hours; eye diseases, namely branch retinal artery occlusion, cilioretinal arteries supplying the macula, combined arterial‐venous occlusion, proliferative diabetic retinopathy, and elevated intraocular pressure (over 30 mmHg); systemic diseases, namely severe general diseases, systemic arterial hypertension (systolic pressure > 200 mmHg), despite therapy, acute systemic inflammation (erythrocyte sedimentation rate > 30 mm within the first hour/C‐reactive protein > 1.0 mg/dL), antithrombin‐III deficiency in case of thrombocytopenia (< 100,000/mL): pathologic clotting time, acute pancreatitis with elevated pancreas enzymes; medical history: heart attack within the last 6 weeks, intracerebral bleeding or neurosurgical operation within the last 4 weeks, therapy with phenprocoumon (Marcumar)/warfarin, allergic reaction to contrast agent, hemorrhagic diathesis, aneurysms, inflammatory vascular diseases (e.g. giant cell arteritis, Wegener's granulomatosis), endocarditis, gastric ulcer; patient participation in other studies during the prior 4 weeks; no willingness and ability of the patient to participate in all follow‐up examinations; pregnancy; written consent not given; patient is not mobile (bedridden); other conditions/circumstances likely to lead to poor treatment adherence (e.g. history of poor compliance, alcohol or drug dependency, no fixed abode) Baseline equivalence: therapy started approximately 2 hours later in the LIF group than in the CST group
Interventions	Local intra‐arterial fibrinolysis (LIF) using recombinant tissue plasminogen activator (rt‐PA): "a microcatheter was superselectively placed into the ophthalmic artery under anticoagulation treatment with heparin (5000 IU) monitored by regular measurements of activated clotting time. In case of occlusion or high‐grade stenosis of the internal carotid artery, the thrombolytic agent was injected into the external carotid artery using the collaterals to the ophthalmic artery. Tissue plasminogen activator (rt‐PA/Actilyse Boehringer Ingelheim Pharm KG, Ingelheim, Germany) was used for thrombolysis. The LIF treatment was terminated after injection of a maximum of 50 mg rt‐PA." Conservative standard treatment (CST): "isovolemic hemodilution was performed according to the hematocrit value. Patients with a hematocrit more than 40% were scheduled for a single isovolemic hemodilution. In these patients, 500 ml of blood was withdrawn and 500 ml of 10% hydroxyethyl starch was simultaneously infused within 15 to 30 minutes. An ocular massage consisting of repeated increased pressure was applied to the globe for 10 to 15 seconds, followed by a sudden release with an in‐and‐out movement using a 3‐mirror contact lens for 3 to 5 minutes. A single eye drop of a topical beta‐blocker (e.g., timolol 0.5%) and intravenous injection of 500 mg acetazolamide were administered to lower intraocular pressure." For both interventions: all participants in the CST and LIF groups were treated with weight‐adapted, low‐dose heparin twice daily for 5 days beginning on the first postinterventional day, and a daily dose of 100 mg of acetylsalicylic acid for at least 4 weeks was initiated
Outcomes	Outcomes reported: change in BCVA 1 month after therapy compared with before therapy; change in visual field using Goldmann’s kinetic perimetry (planned but not assessable due to lack of data)   Adverse outcomes: safety of the 2 procedures in terms of adverse reactions Measurement time points: 1 month Other issues with outcome assessment (e.g. quality control for outcomes, if any): none
Notes	Study period: between 2002 and 2007 Publication language: English Trial registration: NCT00637468 Conflicts of interest: "The author(s) have no proprietary or commercial interest in any materials discussed in this article." Funding source: "The German Research Foundation of Health (DFG), Bonn, Germany, provided financial support (SCHU1454/1‐3). The study does not necessarily reflect the view of the Foundation. Boehringer Ingelheim Pharma KG also provided funding. The Clinical Trials Center, University Hospital Freiburg, received funding from the BMBF (Federal Ministry of Education and Research)."

Incandela 2002.

Study characteristics
Methods	Study design: parallel‐group randomized controlled trial Number randomized (total and per group): 10 eyes of 10 participants in total; 5 eyes of 5 participants in each group Unit of randomization (individual or eye): individual  Number analyzed (total and per group): 10 eyes of 10 participants in total; 5 eyes of 5 participants in each group Unit of analysis (individual or eye): individual (1 eye per participant was included) Exclusions and losses to follow‐up (total and per group): none How were missing data handled?: NA Length of follow‐up: 4 weeks Reported power calculation (Y/N), if yes, sample size and power: not reported
Participants	Country: Italy Setting: university hospital Baseline characteristics: Pentoxifylline, n = 5 Age (mean ± SD, range): 51 ± 4 years Gender: 4 men and 1 woman Time from symptoms to diagnosis/presentation: not reported Time from symptoms to treatment: not reported Placebo, n = 5 Age (mean ± SD, range): 50 ± 6 years Gender: 3 men and 2 women Time from symptoms to diagnosis/presentation: not reported Time from symptoms to treatment: not reported Overall, n = 10 Age (mean ± SD, range): not reported Gender: 7 men and 3 women Time from symptoms to diagnosis/presentation: not reported Time from symptoms to treatment: not reported Inclusion criteria: sudden loss of vision associated with thrombosis of the retinal artery, and decrease in retinal flow and asymmetry between the 2 retinal arterial blood flow (> 40%) documented with duplex scanning Exclusion criteria: previous coronary or vascular surgery, heart or renal failure, diabetes mellitus Baseline equivalence: comparable
Interventions	Intervention: pentoxifylline tablets 600 mg 3 times a day Control: placebo Duration: both for 4 weeks
Outcomes	Outcomes reported: change in central retinal artery flow velocity. Self‐reported clinical improvement measured on an analogue scale line (0 to 10) Adverse outcomes: no adverse events were observed Measurement time points: 4 weeks Other issues with outcome assessment (e.g. quality control for outcomes, if any): none
Notes	Study period: not reported Publication language: English Trial registration: not reported Conflicts of interest: not reported Funding source: not reported

Naycheva 2013.

Study characteristics
Methods	Study design: parallel‐group randomized controlled trial Number randomized (total and per group): 13 participants in total (1 of them was BRAO); 3 participants in 0 mA TES (sham), 5 participants in 66% of the participant's individual EPT at 20 Hz (60%), 5 participants in 150% of the participant's individual EPT at 20 Hz (150%) Unit of randomization (individual or eye): individual  Number analyzed (total and per group): 13 participants in total (1 of them was BRAO); 3 participants in 0 mA TES (sham), 5 participants in 66% of the participant's individual EPT at 20 Hz (60%), 5 participants in 150% of the participant's individual EPT at 20 Hz (150%) Unit of analysis (individual or eye): individual Exclusions and losses to follow‐up (total and per group): none How were missing data handled?: NA Length of follow‐up: 17 weeks Reported power calculation (Y/N), if yes, sample size and power: N
Participants	Country: Germany Setting: single center, university hospital Baseline characteristics: Overall, n = 13 Age (mean ± SD, range): age range 25 to 84 years, median 74 years Gender: not reported Inclusion criteria: age > 18 years and BCVA between light perception and 0.7 logMAR Exclusion criteria: presence of other ophthalmologic diseases, in particular proliferative retinal diseases, such as diabetic retinopathy, retinal or choroidal neovascularization, exudative age‐related macular degeneration, silicone oil tamponade, and severe general diseases Baseline equivalence: not reported
Interventions	Transcorneal electrical stimulation 0 mA Transcorneal electrical stimulation 66% of the participant’s individual electrical phosphene threshold (EPT) at 20 Hz Transcorneal electrical stimulation 150% of the participant’s individual EPT at 20 Hz
Outcomes	Outcomes reported: BCVA, EPT, visual field Adverse outcomes: not reported Measurement time points: 8 follow‐up visits over 17 weeks Other issues with outcome assessment (e.g. quality control for outcomes, if any): none
Notes	Study period: not reported Publication language: English Trial registration: not reported Conflicts of interest: "None" Funding source: "The study was supported financially by Okuvision GmbH, Reutlingen, Germany."

Wang 2017.

Study characteristics
Methods	Study design: parallel‐group randomized controlled trial Number randomized (total and per group): 50 eyes of 50 participants in total; 26 in the ophthalmic artery branch retrograde intervention, 24 in the superselective ophthalmic artery/selective carotid intervention Unit of randomization (individual or eye): individual  Number analyzed (total and per group): 50 eyes of 50 participants in total; 26 in the ophthalmic artery branch retrograde intervention, 24 in the superselective ophthalmic artery/selective carotid intervention Unit of analysis (individual or eye): individual Exclusions and losses to follow‐up (total and per group): none How were missing data handled?: NA Length of follow‐up: 5 to 21 days (mean follow‐up period: 16 days) Reported power calculation (Y/N), if yes, sample size and power: N
Participants	Country: China Setting: tertiary care hospital Baseline characteristics: Ophthalmic artery branch retrograde thrombolytic intervention, n = 26 Age (mean ± SD, range): 56.25 ± 13.23 years Gender: 15 men and 11 women Time from symptoms to diagnosis/presentation: 21.97 ± 13.29 hours Time from symptoms to treatment: not reported Superselective ophthalmic artery thrombolysis/selective carotid intervention, n = 24 Age (mean ± SD, range): 54.33 ± 11.57 years Gender: 14 men and 10 women Time from symptoms to diagnosis/presentation: 21.97 ± 10.13 hours Time from symptoms to treatment: not reported Overall, n = 50 Age (mean ± SD, range): 55.7 ± 2.3 years Gender: 28 men and 22 women Time from symptoms to diagnosis/presentation: 22 hours, range 8 to 76 hours Time from symptoms to treatment: not reported Inclusion criteria: aged between 18 and 78 years, within 76 h after onset, blood pressure below 180/100 mmHg (1 mmHg = 0.133 kPa), and normal prothrombin time, thrombin time, fibrinogen, and platelet counts Exclusion criteria: CT‐confirmed intracranial bleeding; a history of intracranial hemorrhage, intracranial arteriovenous malformation, or aneurysm; intracranial surgery; head trauma or stroke within the past month; active gastrointestinal or urinary bleeding with the past 21 days; surgery within the past 14 days; ultrasound‐confirmed cardiogenic emboli and carotid atherosclerotic plaque; and history of severe heart, lung, kidney, and liver disease Baseline equivalence: comparable
Interventions	Ophthalmic artery branch retrograde thrombolytic intervention: "the skin around the diseased eye was disinfected and the supraorbital nerve was blocked by an injection of 1% lidocaine. The skin below the medial canthus or above the supraorbital foramen was cut, and the supratrochlear or supraorbital artery was exposed by blunt dissection (Figure 1). Distal arterial blood flow was blocked, and the vascular wall was pierced by a sharp knife to insert a microcatheter with a guidewire (Figure 2). The guidewire was removed and 4 mL of the contrast agent iohexol 300 (National Drug registration No. H10970325, Yangtze River Pharmaceutical Group) was injected into the artery a speed of 1 mL/s through the microcatheter using a 1 mL injection syringe. The ophthalmic artery was examined by DSA to confirm the correct insertion of the microcatheter (Figure 3). A total amount of 400,000 U of urokinase and 30 mg of papaverine was pumped into the artery through a coaxial microcatheter using a micropump. The microcatheter was removed and the artery ligated. The surgery was completed after the skin was sutured. The patients were not required to rest in bed after the surgery." Superselective ophthalmic artery thrombolysis/selective carotid intervention: "The femoral artery in the groin area was punctured using the modified Seldinger technology. Briefly, the skin at the puncture site was disinfected, and the body part near the femoral artery was anesthetized by 2% lidocaine. A 2‐mm incision was created, and the femoral artery was pierced by an arterial puncture needle. A guidewire was placed into the femoral artery in the right position, and a 5F arterial catheter sheath, cannula, and catheter were inserted sequentially through the aorta, common carotid, and internal carotid artery to reach the ophthalmic artery. A total amount of 8 mL and 6 mL of the contrast agent iohexol 300 was injected in the internal carotid and the external carotid, respectively, at a speed of 2 mL/s using a high‐pressure syringe. A total amount of 400,000 U of urokinase and 30 mg of papaverine was pumped into the artery through a coaxial microcatheter using a micropump. The catheter was removed and the wound was dressed. The patients were required to rest in bed and closely monitored for bleeding at the puncture site and cardiovascular abnormities for 24 h after the surgery."
Outcomes	Outcomes reported: clinical effects based on the following criteria:  an effective treatment;  the mean arm‐retina circulation time;  a visual acuity test. Adverse outcomes: systemic and local adverse reactions Measurement time points: not reported Other issues with outcome assessment (e.g. quality control for outcomes, if any): none
Notes	Study period: between March 2000 and September 2015 Publication language: English Trial registration: not reported Conflicts of interest: "The author declare no conflict of interests." Funding source: "Source of support: Departmental sources"

Wu 2016.

Study characteristics
Methods	Study design: parallel‐group randomized controlled trial Number randomized (total and per group): 48 eyes of 48 participants in total; 24 participants in each group Unit of randomization (individual or eye): individual  Number analyzed (total and per group): not clearly reported Unit of analysis (individual or eye): individual (1 eye per participant was included) Exclusions and losses to follow‐up (total and per group): not clearly reported How were missing data handled?: not reported Length of follow‐up: average of 3 months Reported power calculation (Y/N), if yes, sample size and power: N
Participants	Country: China Setting: university hospital Baseline characteristics: Recombinant tissue plasminogen activator (rt‐PA) combined with compound anisodine, n = 24 Age (mean ± SD, range): 58.8, range 42 to 76 years Gender: 16 men and 8 women Time from symptoms to diagnosis/presentation: not reported Time from symptoms to treatment: not reported Conventional treatment, n = 24 Age (mean ± SD, range): 50.2, range 36 to 72 years Gender: 14 men and 10 women Time from symptoms to diagnosis/presentation: 21.97 ± 10.13 hours Time from symptoms to treatment: not reported Overall, n = 47 Age (mean ± SD, range): not reported Gender: not reported Time from symptoms to diagnosis/presentation: not reported Time from symptoms to treatment: not reported Inclusion criteria: CRAO Exclusion criteria: tendency of active hemorrhage and known tendency of bleeding; a history of cerebral infarction or myocardial infarction in recent 6 months; current adoption of anticoagulation; intracranial arterial aneurysm or arteriovenous malformation; serious heart failure and dysfunction of liver and kidney; thrombocyte count < 100 x 10000000000/L; and gestation and non‐co‐operation Baseline equivalence: comparable
Interventions	Recombinant tissue plasminogen activator combined with compound anisodine (group 1): "0.5 mg of sublingual nitroglycerin tablets immediately melted 3 times a day. Once per day, retrobulbarly was injected with 0.5 ml of atropine, and was orally administered with 50 mg of methazolamide tablets twice a day. The cases inhaled oxygen (the gas mixture of oxygen with a volume fraction of 95% and carbon dioxide with a volume fraction of 5%) for 10 min every 1 hour, eyeballs were massaged repetitively and intermittently, alternatively eyeballs were pressed on the patients' upper palpebras with index fingers of both hands for 10‐15 sec each time" Recombinant tissue plasminogen activator combined with compound anisodine (group 2): "rt‐PA within 30 mins (0.5 mg·kg‐1; the maximum dosage being 90 mg), initially 10% of the dose was injected through veins and the 90% via intravenous drip within 1 hour. The signs were monitored and observed for mucocutaneous, respiratory and archenteric hemorrhage or bleeding in the urinary tract every 0.5 h within 3 h after thrombolysis, and then monitored once every hour within 24 hours. Anticoagulation and antiplatelet agents were forbidden within 24 h. If no hemorrhage occurred, 80 mg of sodium ozagrel was given and 100 ml of 9 g·l‐1 (10‐13). Sodium chloride injection was added, which was adopted twice a day through intravenous drip. If any tendency of hemorrhage was identified, the medicine was immediately stopped and antifibrinolytic agents given, and fresh plasma or blood was transfused if necessary." Recombinant tissue plasminogen activator combined with compound anisodine (group 3): "these cases were simultaneously injected with 2 ml of compound anisodine subcutaneously beside the lateral superficial temporal artery twice a day, and continued for 2 courses of treatment with 14 days being one course" Recombinant tissue plasminogen activator combined with compound anisodine (group 4): "these cases were orally administered with 0.5 g of mecobalamin tablets three times a day and injected intramuscularly 0.1 g vitamin B1 and 0.5 mg of B12. Days (13 or 14) after medications, visual acuity, fundus oculi, and visual field were monitored; hemostasis and hepatorenal function were respectively checked 24 h and 7 days after the treatment. The treatment in the treatment group lasted 14‐52 days, with an average period of 38 days" Conventional treatment (group 1): "immediate eyeball massage as well as oxygen inhalation was given, and patients were asked to have melted sublingual nitroglycerin. Retrobulbarly, 0.5 ml of atropine was injected once a day, and 50 mg of methazolamide tablets were orally administered twice a day. The course of treatment lasted for 1 week." Conventional treatment (group 2): these cases were given 0.4 g of puerarin injection within half an hour and intravenous drip of 250 ml of 50 g·l‐1 glucose injection or 250 ml of 9 g·l‐1 sodium chloride injection was provided once a day, and one course of treatment lasted for 2 weeks Conventional treatment (group 3): "these cases were simultaneously injected with 2 ml of compound anisodine subcutaneously beside the lateral superficial temporal artery twice a day for 2 courses of treatment with one course being 14 days." Conventional treatment (group 4): "these cases were orally administered with 0.1 g of aspirin enteric‐coated tablets once a day, orally administered 0.5 g of mecobalamine tablets three times a day, intramuscularly injected with 0.1 g of vitamins B1 and B12. The treatment in the control group lasted for 15‐58 days, with an average period of 39 days"
Outcomes	Outcomes reported: visual acuity, fundus oculi and visual field to monitor fibrinogen changes and observe curative effects Adverse outcomes: gum bleeding after being given intravenous thrombolysis Measurement time points: NA Other issues with outcome assessment (e.g. quality control for outcomes, if any): none
Notes	Study period: June 2009 to June 2012 Publication language: English Trial registration: not reported Conflicts of interest: not reported Funding source: not reported

BCVA: best‐corrected visual acuity

BRAO: branch retinal artery occlusion

CRAO: central retinal artery occlusion

CST: conservative standard treatment

CT: computed tomography

EECP: enhanced external counterpulsation 

EPT: electrical phosphene threshold

INR: international normalized ratio

LIF: local intra‐arterial fibrinolysis

logMAR: logarithm of the minimum angle of resolution

NA: not applicable

Nd:YAG: neodymium‐doped yttrium aluminum garnet

OCT: optical coherence tomography

rt‐PA: recombinant tissue plasminogen activator

SD: standard deviation

TES: transcorneal electrical stimulation

t‐PA: tissue plasminogen activator 

VA: visual acuity

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
ACTRN12609001042235	Not study design of interest: non‐randomized study
Arnold 2005	Retrospective case‐control study. 37 patients were treated with intra‐arterial urokinase within 6 hours of onset of symptoms, and were also given aspirin. 19 patients who also sustained a CRAO within 6 hours but did not have thrombolysis (for a variety of patient and non‐patient reasons) were used as controls and were given aspirin or heparin. Some cases and some controls had lowering of their IOP by either paracentesis or intravenous acetazolamide (Diamox)—the decision to give this was "according to the preference of the treating ophthalmologist".
Atebara 1995	Comparison of paracentesis with carbogen (95% oxygen and 5% carbon dioxide). Retrospective, case‐notes review
Augsburger 1980	32 patients with CRAO. Standard treatment of paracentesis, massage, carbogen, acetazolamide, and aspirin. No control group
Bedersdorfer 2022	Not study design of interest: case report
Callizo 2009	Not study design of interest: non‐randomized study
ChiCTR2000038317	Not study design of interest: case‐control study
Drago 1982	Not study design of interest: not a trial
Fernández 2000	Not study design of interest: non‐randomized study
Gu 1999	Not study design of interest: not a trial
Hayreh 2008	Not study design of interest: not a trial
Holschermann 2005	Pilot non‐controlled study. Used tirofiban (intravenous platelet inhibitor) in 18 consecutive patients
Iwafune 1980	Compared pentoxifylline with various "conventional fibrinolytic agents". Only 2 CRAO cases and 3 controls. Non‐randomized, non‐standardized treatment of controls
JPRN‐jRCT2031210280	Not participants of interest: healthy male
JPRN‐jRCT2032190199	Not study design of interest: no control group
JPRN‐jRCTs031180441	Not study design of interest: no control group
JPRN‐jRCTs041190079	Not study design of interest: no control group
JPRN‐UMIN000012161	Not study design of interest: no control group
JPRN‐UMIN000022972	Not study design of interest: no control group
JPRN‐UMIN000023979	Not study design of interest: no control group
JPRN‐UMIN000036219	Not study design of interest: no control group
Kattah 2002	Case series of 12 patients with CRAO. Patients who presented with IOP of more than 12 mmHg had anterior chamber paracentesis initially. All patients were given intra‐arterial rt‐PA, which was followed by heparin and then warfarin for 1 month. The latter was replaced with aspirin or clopidogrel after 1 month.
Makhkamova 2015	Not participants of interest: not participants with CRAO
Mehboob 2021	This study included participants with hemiretinal artery or major branch retinal artery.
Mendes 2018	Not study design of interest: case report
Neubauer 2000	Retrospective uncontrolled review of a number of non‐invasive treatments for CRAO
Noble 2009	Not study design of interest: not a trial
Pettersen 2005	Retrospective notes review of 8 cases. All given rt‐PA intra‐arterially.
Richard 1999	46 patients with CRAO underwent local intra‐arterial fibrinolysis with rt‐PA. Retrospective study with no control group
Rumelt 1999	11 patients with CRAO were treated with an "aggressive" systematic regimen until retinal circulation improved or treatment modalities were exhausted. The results were compared with 5 patients who were treated in an "arbitrary non‐systematic manner". There was thus no true control group, and multiple interventions were compared.
Schmidt 1992	14 consecutive patients with CRAO treated with urokinase via the ophthalmic artery. Confounded, as patients were then heparinized for 3 days. Control group consisted of 41 consecutive patients, but there is no mention of how these patients were chosen.
Schumacher 1993	23 patients with CRAO underwent local intra‐arterial fibrinolysis. No control group
Tan 2009	Not intervention of interest
TCTR20210510007	Not study design of interest: observational study
Telek 2019	Not study design of interest: retrospective study
Weber 1998	Retrospective and non‐randomized. Ophthalmic artery urokinase given to 17 patients compared with 15 patients who did not receive fibrinolysis. In both groups some of the patients received other treatments such as paracentesis.
Werner 2004	This study included participants with CRAO and branch retinal arterial occlusion, and the results were not separately reported for those heterogeneous populations.

CRAO: central retinal artery occlusion
IOP: intraocular pressure
rt‐PA: recombinant tissue plasminogen activator
 

Characteristics of ongoing studies [ordered by study ID]

JPRN‐jRCT2021190013.

Study name	Phase IIa study (doctor‐led trial) for the efficacy and safety of SJP‐0008 in central retinal artery occlusion
Methods	Parallel‐group randomized controlled trial
Participants	Number randomized: 20 (planned) Inclusion criteria: Age 20 years or older at the time consent was obtained  Written consent for joining the study was available  The onset of non‐arteritic retinal artery occlusion (central retinal artery occlusion) occurred expected to start study drug administration within 3 to 48 hours before joining the study  Decimal visual acuity in the included eye better than hand motion and worse than 0.1 in a screening test All of the following in the screening test before the study: Hemoglobin amount >= 10 g/dL  White blood cell count >= 2000/uL  Platelet count >= 100,000/uL  AST =< 50 IU/L  ALT =< 50 IU/L  Total bilirubin =< 1.5 mg/dL  Creatinine =< 1.5 mg/dL  Calcium >= 7.5 mg/dL, =< 10.5 mg/dL Exclusion criteria: Presence of vision‐affecting disease in the included eye at risk of progression during the trial, including diabetic retinopathy, retinal detachment, macular disease, or retinitis pigmentosa. However, such patients were included if the primary or other physician deemed inclusion unlikely to affect the safety of the patient.  History of internal eye surgery (including laser treatment) within 90 days before enrollment Patients with less than 0.1 visual acuity  From the time of obtaining consent to the end of the study drug administration period, patients who are not willing to perform appropriate contraception using effective contraceptive methods (e.g. condoms, intrauterine devices, oral contraceptives), pregnant women or lactating women  Patients undergoing unapproved or off‐label treatment in other clinical trials within 4 weeks before the time consent was obtained  Patients who had received the study drug in the past  Patients with poor control of comorbidity  A judgement that inclusion was inappropriate by the primary or other physician for any other reason
Interventions	Intervention: SJP‐0008 200 mg/day (100 mg test drug 2 tablets) once daily after breakfast for up to 29 days  Control: SJP‐0008 100 mg/day (100 mg test drug and 1 placebo tablet) once daily after breakfast for up to 29 days
Outcomes	Primary outcome: Change in ETDRS visual acuity at the end of the observation period Secondary outcomes: Change in ETDRS visual acuity in both eyes Improvement and deterioration of ETDRS vision in both eyes Retinal visibility, central vision in target eye Fluorescein angiography in target eye Laser speckle blood flow test in both eyes Optical coherence tomography in both eyes
Starting date	Started in December 2019
Contact information	Dr Nakazawa Toru, ntoru@oph.med.tohoku.ac.jp
Notes

JPRN‐jRCTs031180280.

Study name	The efficacy of hydrogen eye drop for the ischemia‐reperfusion injury of retinal artery occlusion
Methods	Parallel‐group randomized controlled trial
Participants	Number randomized: 30 participants (planned) Inclusion criteria: patient of retinal arterial occlusion, aged > 20 years Exclusion criteria: patients without thrombolytic therapy
Interventions	Intervention: hydrogen eye drop will be applied to the eye (50 mL/min, 20 min, total 1000 mL) Control: placebo
Outcomes	Primary outcome: visual acuity, retinal optical coherence tomography Secondary outcomes: arm‐to‐retina circulation time by fluorescein angiography, static visual field, retinal electrogram
Starting date	Started in November 2014
Contact information	Dr Tsutomu Igarashi, tutomu@nms.ac.jp
Notes

NCT03197194.

Study name	A phase III randomized, blind, double dummy, multicenter study assessing the efficacy and safety of intravenous thrombolysis (Alteplase) in patients with acute central retInal artery occlusion
Methods	Parallel‐group randomized controlled trial
Participants	Number randomized: 70 participants (planned) Inclusion criteria: Patients aged up to 18 years CRAO diagnosis by fundoscopic examination or non‐mydriatic retinophotography (NMR) performed by an ophthalmologist Blindness defined according to WHO classification as visual acuity < 1/20 (20/400) Treatment intervention should be initiated by a stroke team as quickly as possible and within 4.5 hours from symptom onset No clinical (e.g headache with jaw claudication or scalp tenderness, no temporal pulse) or laboratory evidence (elevated CRP) of giant cell arteritis No clinical or radiological evidence of stroke within the last 3 months Patients covered by healthcare insurance (social security) Written informed consent obtained Exclusion criteria: Symptoms onset more than 4.5 hour prior to infusion start or undetermined time of symptom onset Minor VA deficit or VA rapidly improving before start of infusion CRAO without foveal ischemia Other retinal vascular disease: occlusion of branch of the CRA without significant VA loss, occlusion of the retinal vein, proliferative diabetic retinopathy or any other severe retinopathy Clinical or laboratory evidence of temporal arteritis. Evidence of ICH or ischemic stroke on the pre‐administration CT scan or MRI Pregnant or lactating women Minors Adults under guardianship or trusteeship Any contraindication to alteplase Any contraindication to aspirin
Interventions	Intervention: intravenous injection of alteplase and 1 placebo tablet Control: 1 tablet of acetylsalicylic acid and 1 dose of intravenous placebo
Outcomes	Primary outcome: VA improvement after treatment at 1 month Secondary outcomes: Tolerance at 3 months Proportion of blindness patients after treatment at 1 month Visual field at 3 months  Time course of VA on ETDRS chart or ordinal scale at 3 months Time‐to‐treatment administration impact on VA evolution at 1 month Global disability (modified Rankin scale) after treatment at 3 months Quality of life related to vision (NEI‐VFQ‐25) after treatment at 3 months
Starting date	Started on 8 June 2018; estimated completion in September 2022
Contact information	Principal Investigator: Dr Benoit Guillon, benoit.guillon@chu‐nantes.fr
Notes

NCT04526951.

Study name	TENecteplase in Central Retinal Artery Occlusion Study (TenCRAOS)
Methods	Parallel‐group randomized controlled trial
Participants	Number randomized: 78 participants (planned) Inclusion criteria: Non‐arteritic central retinal artery occlusion with ≥ 1.0 logMAR visual acuitiy and symptoms lasting less than 4.5 hours Ability to administer the Investigator Medicinal Product (IMP) within 4.5 hours of symptom onset Age ≥ 18 years Informed written consent of the patient A woman of childbearing potential must confirm that in her opinion, she cannot be pregnant, or if there is a possibility that she is pregnant, a negative pregnancy test must be confirmed before any IMP is given Exclusion criteria:  No other active intervention targeting CRAO Branch retinal artery occlusion, cilioretinal artery supplying the macula, combined arterial‐venous occlusion, proliferative diabetic retinopathy, elevated intraocular pressure (> 30 mmHg) or clinical suspicion of ophthalmic artery occlusion occlusion (e.g. choroidal non‐perfusion, absence of cherry red spot, no light perception) Systemic diseases; severe general diseases, systemic arterial hypertension (blood pressure > 185/110 mmHg), despite medical therapy, or clinical suspicion of acute systemic inflammation Presence of intracranial hemorrhage on brain MRI/CT Medical history: heart attack within the last 6 weeks, intracerebral bleeding or neurosurgical operation within the last 4 weeks, therapy with anticoagulation, allergic reaction to contrast agent, hemorrhagic diathesis, aneurysms, inflammatory vascular diseases (e.g. giant cell arteritis, granulomatosis with polyangitis), endocarditis, or gastric ulcer No willingness and ability of the patient to participate in all follow‐up examinations Pregnancy (if suspicion of pregnancy s‐hCG or u‐hCG must be negative) Allergy or intolerance to any ingredients of IMP or placebo or gentamicin Other conditions/circumstances likely to lead to poor treatment adherence (e.g. history of poor compliance, alcohol or drug dependency, no fixed abode) Significant bleeding disorder either at present or within the past 6 months Effective oral anticoagulant treatment, e.g. warfarin sodium (INR > 1.3) Effective anticoagulant treatment with heparin or low molecular weight heparin the last 48 hours Any history of central nervous system damage (i.e. neoplasm, aneurysm, intracranial or spinal surgery) Known hemorrhagic diathesis Major surgery, biopsy of a parenchymal organ, or significant trauma within the past 2 months (including any trauma associated with acute myocardial infarction) Recent non‐compressible vessel puncture within 2 weeks Recent trauma to the head or cranium Prolonged cardiopulmonary resuscitation (> 2 minutes) within the past 2 weeks Acute pericarditis or subacute bacterial endocarditis, or both Acute pancreatitis Severe hepatic dysfunction, including hepatic failure, cirrhosis, portal hypertension (esophageal varices), and active hepatitis Active peptic ulceration Arterial aneurysm and known arterial/venous malformation Neoplasm with increased bleeding risk Any known history of hemorrhagic stroke or stroke of unknown origin Known history of ischemic stroke or transient ischemic attack in the preceding 3 months Dementia
Interventions	Intervention: intravenous injection of tenecteplase and 1 dose of placebo tablet Control: 1 tablet of acetylsalicylic acid and 1 dose of intravenous placebo
Outcomes	Primary outcome: proportion of participants with ≤ 0.7 logMAR visual acuity in the affected eye at 30 (±5) days after treatment Secondary outcomes (timeframe 30 (±5) and 90 (±15) days unless specified): Proportion of participants with ≤ 0.5 logMAR visual acuity in the affected eye Mean improvement in logMAR visual acuity in the affected eye Proportion of participants with visual recovery (logMAR ≤ 0.7) and (logMAR ≤ 0.5) in the affected eye Number of test points seen (of 100) on monocular Esterman perimetry Acute ischemic lesions on follow‐up on diffusion‐weighted (DWI) MRI or on brain CT at baseline and 24 h NIHSS score at 24 h and discharge Modified Rankin scale score at discharge Mean score on NEI‐VFQ‐25  Mean score on EQ‐5D Presence of ocular neovascularization Other outcomes (timeframe 30 (±5) and 90 (±15) days unless specified): All‐cause and stroke‐related death at discharge Proportion of participants with any intracranial hemorrhage at 24 h Proportion of participants with symptomatic intracranial hemorrhage until discharge Proportion of participants with complications such as systemic bleeding at 24 h, discharge, and 30 (±5) days Other serious adverse events Occurrence of adverse events Retrobulbar spot sign detection using duplex/Doppler ultrasound at baseline, point‐of‐care ultrasound Retrobulbar spot sign and central retinal artery recanalization using duplex/Doppler ultrasound at 24 h and discharge Macular optical coherence tomography (OCT) volume scans and macular OCT angiography
Starting date	Started on 30 October 2020; estimated completion date 31 January 2024
Contact information	Senior consultant: Dr Anne Hege Aamodt,  a.h.aamodt@medisin.uio.no
Notes

REVISION.

Study name	REVISION trial
Methods	Parallel‐group randomized controlled trial
Participants	Number randomized: 1400 participants (planned) Inclusion criteria: Acute non‐arteritic CRAO (i.e. sudden, painless monocular vision loss) ≤ 12 hours after symptom onset confirmed by an experienced ophthalmologist through assessment of: BCVA, intraocular pressure, swinging flash light test (relative afferent pupil defect), slit‐lamp biomicroscopy, fundoscopy, and OCT of the macula of both eyes* (*within the 4.5‐hour time window: to be skipped if not feasible ≤ 10 minutes; beyond the 4.5‐hour time window: mandatory) BCVA of logMAR ≥ 1.3 in the affected eye (functional blindness according to WHO ICD‐11) Reading must have been possible with the affected eye before CRAO (logMAR ≤ 0.5) Neurological examination performed by an experienced stroke neurologist Brain imaging as per local standard for acute retinal ischemia/stroke assessment, either cranial CT or cranial MRI Exclusion criteria:  Suspected giant cell arteritis Other‐than‐CRAO cause of acute visual loss (e.g. retinal detachment, vitreous hemorrhage, acute glaucoma, acute optic neuritis) BCVA of logMAR < 1.3 or rapidly improving vision in the affected eye Acute ischemic stroke with indication for on‐label intravenous thrombolysis Any co‐existing or terminal disease with anticipated life expectancy of < 3 months Prior participation in the REVISION trial
Interventions	Intervention: thrombolysis  Alteplase (0.9 mg per kg body weight; 10% as bolus; remaining over 1 hour) will be administered intravenously within 4.5 hours of symptom onset. Control: placebo  Placebo (0.9 mg per kg body weight; 10% as bolus; remaining over 1 hour) will be administered intravenously within 4.5 hours of symptom onset.
Outcomes	Primary outcome: functional recovery at 30 days Secondary outcomes (timeframe 90 days unless specified): BCVA Shift in visual outcome categories Dichotomized analysis of visual outcome Visual field Central retinal artery recanalization Retinal arterial perfusion NEI‐VFQ‐25 NIHSS score Modified Rankin scale score Fresh ischemic lesions on cranial MRI Death  Any ICH within 72 hours Symptomatic ICH within 72 hours Intraocular hemorrhage in the affected eye within 72 hours Major bleeding within 72 hours Retinal neovascularization requiring therapy Serious adverse events
Starting date	Started on 1 March 2022; estimated completion date 31 December 2025
Contact information	Dr Maximilian Schultheiss, m.schultheiss@uke.de
Notes

ALT: alanine aminotransferase

AST: aspartate transaminase

BCVA: best‐corrected visual acuity

CRA: central retinal artery

CRAO: central retinal artery occlusion

CRP: C‐reactive protein

CT: computed tomography

ETDRS: Early Treatment Diabetic Retinopathy Study 

hCG: human chorionic gonadotropin

ICD: International Classification of Diseases

ICH: intracranial hemorrhage

IMP: investigator medicinal product

INR: international normalized ratio

IU: international units

logMAR: logarithm of the minimum angle of resolution 

MRI: magnetic resonance imaging

NEI‐VFQ‐25: National Eye Institute Visual Function Questionnaire 

NIHSS: National Institutes of Health Stroke Scale 

OCT: optical coherence tomography

VA: visual acuity

WHO: World Health Organization

Differences between protocol and review

2009 update

The intervention enhanced external counterpulsation (EECP) was added to the inclusion criteria of the review.

2023 update

We updated the Methods to follow current Cochrane methodology. 

Contributions of authors

Screening search results: SS, JL
Assessing trials according to inclusion criteria: JL, SN, SS
Assessing quality of trials: SS, SN
Extracting data: JL, SN
Writing the text of the review: JL, SS, PG
Reviewing and revising the review: JL, SS, SN, IS, PG

Sources of support

Internal sources

• None, Other

  No internal source of support.

External sources

• Cochrane Eyes and Vision US Project, USA

  Supported by grant UG1EY020522 (PI: Tianjing Li, MD, MHS, PhD), National Eye Institute, National Institutes of Health

• Queen's University Belfast, UK

  Gianni Virgili, Co‐ordinating Editor for Cochrane Eyes and Vision’s work is funded by the Centre for Public Health, Queen’s University of Belfast, Northern Ireland 

Declarations of interest

John C Lin: none known
Sophia Song: none known
Sueko M Ng: none known
Ingrid U Scott: none known
Paul B Greenberg: none known

New search for studies and content updated (no change to conclusions)