← Back

CRAO AGILE Metanalysis

Early Thrombolysis and Outcomes in Central Retinal Artery Occlusion: An Individual Participant Data Meta-Analysis

Share Share Copied! Compare

Year of Publication: 2025

Authors: Jim S. Xie, MD, et al, for the Assessment Group for Interventional Lysis in Eye (AGILE)

Journal: Stroke

Citation: Stroke. 2025:56:00-00. DOI: 10.1161/STROKEAHA.124.049955

Link: https://www.ahajournals.org/doi/suppl/10...EAHA.124.049955

PDF: https://www.ahajournals.org/doi/epub/10....EAHA.124.049955

Clinical Question

To determine whether time to treatment influences the effect of intraarterial thrombolysis (IAT), intravenous thrombolysis (IVT), and conservative standard therapy (CST) on visual outcomes in nonarteritic central retinal artery occlusion (CRAO).

Bottom Line

Early intervention in nonarteritic central retinal artery occlusion is associated with better visual recovery, with intraarterial thrombolysis (IAT) and intravenous thrombolysis (IVT) outperforming nonthrombolytic treatments.

        Major Points
        This was an individual participant data (IPD) meta-analysis of 35 studies, including data from 783 patients with nonarteritic central retinal artery occlusion (CRAO).
For every hour decrease in time to treatment, thrombolysis was associated with greater improvement in best-corrected visual acuity (BCVA) than conservative standard therapy.
Intraarterial thrombolysis (IAT) given within 6 hours was associated with an increased odds of recovery from severe vision loss (OR, 2.72) compared to conservative standard therapy.
Intravenous thrombolysis (IVT) given within 4.5 hours was associated with an increased odds of recovery from severe vision loss (OR, 3.32) compared to conservative standard therapy.
A changepoint for IAT efficacy was detected at 8 hours, suggesting a potential treatment window beyond current guidelines.
The findings of this meta-analysis warrant confirmation in sufficiently powered randomized controlled trials.

      

Design

Study Type: Individual participant data meta-analysis.

Randomization:

Enrollment Period: Studies up to June 2023 were included.

Follow-up Duration: Follow-up time did not differ between groups.

Countries: United States, Canada, Germany, South Korea, Australia, Switzerland, France, Israel, Austria

Sample Size: 783

Analysis: Mixed-effect models, local polynomial regression, changepoint analysis, and Monte Carlo simulations.

Inclusion Criteria

Participants presenting with severe vision loss (final BCVA <1.0 logMAR).
Treated within 24 hours of symptom onset.
Diagnosis of nonarteritic central retinal artery occlusion (CRAO).
Original studies reporting on visual outcomes of at least 5 individuals.

Exclusion Criteria

Patients with nonsevere vision loss (BCVA <1.0 logMAR).
Patients treated after 24 hours of symptom onset.
Studies where CRAO was associated with an additional retinal vascular occlusion or a macula-supplying cilioretinal artery.
CRAO secondary to giant cell arteritis, other autoimmune vasculitis, trauma, or an iatrogenic source.
Studies of prostaglandin E1 were excluded.

Baseline Characteristics

Characteristic	Control	Active
Age	65.6 (SD 13.0) years.	IAT: 62.5 (SD 13.9) years; IVT: 67.7 (SD 11.9) years.
Female	77 (34.2%).	IAT: 107 (35.7%); IVT: 62 (36.9%).
Pretreatment BCVA	2.2 (SD 0.5) logMAR.	IAT: 2.3 (SD 0.5) logMAR; IVT: 2.4 (SD 0.4) logMAR.
Treatment window <=4.5 h	45 (18.9%).	IAT: 30 (8.5%); IVT: 132 (69.1%).
Treatment window 4.5-6 h	38 (16.0%).	IAT: 95 (26.8%); IVT: 31 (16.2%).
Treatment window 6-12 h	76 (31.9%).	IAT: 159 (44.9%); IVT: 18 (9.4%).
Treatment window 12-24 h	79 (33.2%).	IAT: 70 (19.8%); IVT: 10 (5.2%).

Arms

Field	Intraarterial Thrombolysis (IAT)	Intravenous Thrombolysis (IVT)	Control
Intervention	Intraarterial administration of thrombolytic agents (including urokinase, streptokinase, and rtPA).	Intravenous administration of thrombolytic agents (including rtPA).	Nonthrombolytic treatments, including ocular massage, anterior chamber paracentesis, hyperbaric oxygen therapy, isovolemic hemodilution, heparin, topical IOP-lowering agents, acetazolamide, corticosteroids, pentoxifylline, aspirin, anticoagulation, and intravenous saline.
Duration	Treatment window of up to 24 hours from symptom onset.	Treatment window of up to 24 hours from symptom onset.	Treatment window of up to 24 hours from symptom onset.

Outcomes

Outcome	Type	Control	Intervention	HR / OR / RR	P-value
Recovery from severe vision loss (rSVL), defined as final best-corrected visual acuity (BCVA) better than 20/200 (<1.0 logMAR).	Primary	12.0% for CST.	27.2% for IAT within 6 hours; 28.8% for IVT within 4.5 hours.		IAT vs CST (within 6 h): 0.05; IVT vs CST (within 4.5 h): 0.02.
Change in best-corrected visual acuity (BCVA) per hour decrease in time to treatment.	Secondary	0.01 logMAR improvement.	0.02 logMAR improvement for IAT; 0.04 logMAR improvement for IVT.		IAT: 0.04; IVT: 0.04; CST: 0.05.
Recovery from functional vision loss (rFVL, final BCVA <=0.7 logMAR).	Secondary		IAT vs CST (within 6 h): OR 3.12 (95% CI, 1.06-9.14); IVT vs CST (within 4.5 h): OR 3.56 (95% CI, 1.06-12.02).	IAT vs CST: 3.12; IVT vs CST: 3.56.	IAT vs CST: 0.04; IVT vs CST: 0.04.
Major complications	Adverse	4 (1.7%).	IAT: 5 (1.4%); IVT: 6 (3.1%).
Minor complications	Adverse	8 (3.4%).	IAT: 13 (3.7%); IVT: 1 (0.5%).

Subgroup Analysis

Subgroup analysis restricted to comparative studies was underpowered and could not reproduce the findings from the main analysis. A subgroup analysis of patients receiving recombinant tissue-type plasminogen activator (rtPA) showed findings and effect sizes similar to the main analysis.

Criticisms

The meta-analysis was primarily based on single-armed studies, which are subject to inherent selection bias and confounding.
The included studies differed in methodological design, eligibility criteria, and treatment protocols, leading to heterogeneity in the IVT and CST groups.
The use of best-corrected visual acuity (BCVA) may not be the most sensitive measure for evaluating treatment efficacy.
Subgroup analysis of comparative studies was underpowered.

Funding

None.

Based on: CRAO AGILE Metanalysis (Stroke, 2025)