PDF: EAGLE
(2010)Objective
To compare treatment outcome of conservative treatment versus intra-arterial tPA in patients with central retinal artery occlusion (CRAO)
Study Summary
• Study stopped early after first interim analysis due to similar efficacy but higher adverse reaction rate in LIF group (37.1% vs 4.3%)
Intervention
CST: Multimodal therapy including isovolemic hemodilution (if hematocrit >40%), ocular massage, topical beta-blocker, IV acetazolamide 500mg, low-dose heparin, aspirin 100mg daily. LIF: Superselective intra-arterial rtPA (Actilyse) up to 50mg via microcatheter into ophthalmic artery under heparin anticoagulation, plus low-dose heparin and aspirin 100mg daily
Inclusion Criteria
Age 18-75 years, non-arteritic CRAO with symptoms ≤20 hours, BCVA worse than 0.5 logMAR (Snellen <20/63) on ETDRS chart
Study Design
Arms: CST group (n=40) vs LIF group (n=44); randomized 1:1 stratified by center using sealed envelopes
Patients per Arm: 40 CST, 44 LIF randomized; 40 CST, 42 LIF in full analysis set; 30 CST, 31 LIF in per-protocol population
Outcome
• Clinically significant improvement (≥0.3 logMAR): 60% CST vs 57.1% LIF
• Final BCVA ≤1.0 logMAR: 15% CST vs 16.7% LIF
• Serious adverse reactions: 2/47 CST (4.3%) vs 13/35 LIF (37.1%) including 2 cerebral/cerebellar hemorrhages in LIF group
• Study terminated early after interim analysis
Bottom Line
Local intra-arterial fibrinolysis with rtPA did not improve visual acuity beyond conservative standard treatment in acute non-arteritic CRAO, despite both treatments showing significant improvement from baseline. Given similar efficacy and higher adverse reaction rates with LIF (37.1% vs 4.3%), the authors cannot recommend LIF for management of acute CRAO.
Major Points
- First prospective randomized multicenter trial comparing LIF with CST in acute CRAO
- 9 centers in Austria and Germany enrolled 84 patients between 2002-2007
- Mean symptom-to-treatment time: 10.99h (CST) vs 12.78h (LIF)
- Both groups showed significant and similar visual improvement at 1 month
- Mean BCVA improvement: -0.44 logMAR (CST) vs -0.45 logMAR (LIF), p=0.69
- Clinically significant improvement (≥0.3 logMAR): 60% CST vs 57.1% LIF
- Study terminated early after first interim analysis on DSMB recommendation
- Conditional power analysis showed only 8.2% probability of detecting significant difference with full 200 patients
- Higher adverse reaction rate in LIF: 37.1% vs 4.3% in CST
- Two serious bleeding events in LIF group: cerebral and cerebellar hemorrhage (both recovered)
- Visual improvement occurred even with longer occlusion times (up to 20-23 hours)
- No difference in treatment effect based on symptom duration (<12h vs ≥12h)
- Independent DSMB oversight with formal interim analyses
Study Design
- Study Type
- Prospective randomized controlled multicenter clinical superiority trial
- Randomization
- Yes
- Blinding
- Open-label (treating physicians unblinded). Endpoint committee retrospectively reviewed eligibility in masked manner based on baseline data, visual fields, fundus photos, and fluorescein angiograms.
- Sample Size
- 84
- Follow-up
- 1 month primary endpoint; up to 1 month
- Centers
- 9
- Countries
- Austria, Germany
Primary Outcome
Definition: Change in best-corrected visual acuity (BCVA) from baseline to 1 month, measured as difference from baseline logMAR using ETDRS charts. Clinically significant improvement defined as decrease in logMAR of ≥0.3 (equivalent to ≥3 Snellen lines). Numeric scores for profound low vision (hand motion, light perception, no light perception) substituted for logMAR values.
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| Mean change -0.443 (SD 0.549); Baseline 2.11 (0.49) to 1-month 1.67 (0.62); 60.0% achieved ≥0.3 logMAR improvement | Mean change -0.447 (SD 0.545); Baseline 2.18 (0.54) to 1-month 1.74 (0.80); 57.1% achieved ≥0.3 logMAR improvement | - (Difference CST vs LIF: 0.004 (SD 0.547); 95% CI -0.236 to 0.245) | 0.69 (Wilcoxon test) |
Limitations & Criticisms
- Study terminated early after first interim analysis - only 84 of planned 200 patients enrolled
- Low conditional power (8.2%) to detect significant difference if study continued
- Open-label design - treating physicians not blinded to treatment allocation
- Mean time to treatment approximately 2 hours longer in LIF group (potential bias)
- 9 patients (11%) had protocol violations: 6 inclusion/exclusion violations, 7 treatment deviations
- 4 patients randomized to LIF received CST due to technical failure
- 3 patients received CST instead of LIF for unknown reasons (randomization failure)
- Primary endpoint missing in 5 patients (3 CST, 2 LIF)
- Recruitment challenges: only 84 of 200 planned patients over 5 years
- Limited generalizability: restricted to patients treatable within 20 hours
- Principal investigator's center showed trend toward better LIF outcomes, raising questions about operator experience
- Visual field secondary outcome not assessable due to lack of data
- No placebo control - both groups received active multimodal treatment
- CST more extensive than in previous trials - may explain better outcomes
- Heparin and aspirin used in both groups (standard in fibrinolysis) - may have improved CST outcomes
- Sample size based on categorized outcome; continuous analysis may have different power
- No anterior chamber paracentesis in CST group (commonly used by some clinicians)
Citation
Ophthalmology 2010;117:1367–1375