Clot-buster fails to restore vision after central retinal artery occlusion

A central retinal artery occlusion is, in effect, a stroke of the eye: the vessel that feeds the retina is blocked, usually by a clot, and vision in that eye can vanish within minutes. By analogy with brain stroke, where clot-busting drugs given quickly can rescue tissue, ophthalmologists have long hoped that the same drugs might rescue sight if given fast enough. A new randomised trial tested that hope directly, and the answer was no. In the phase 3 TenCRAOS study, intravenous tenecteplase given within 4.5 hours of symptom onset did not restore vision any better than aspirin, and it caused more bleeding [s1].

The appeal of the idea is easy to understand. Effective treatment for central retinal artery occlusion is lacking, and the loss is often permanent [s1]. Tenecteplase is the same clot-dissolving drug now used routinely in acute ischaemic stroke of the brain, where a narrow treatment window separates recovery from disability. Extending it to the eye seemed a reasonable bet — but reasonable bets in medicine still have to be tested against a control, because the natural history of any condition, and the placebo-like effect of intensive care, can flatter a treatment that does nothing.

What they did

TenCRAOS was a phase 3, double-blind, double-dummy, randomised, controlled trial [s1]. It enrolled adults with acute, non-arteritic central retinal artery occlusion whose symptoms had begun within 4.5 hours before treatment [s1]. Participants were assigned in a 1:1 ratio to one of two regimens: intravenous tenecteplase at 0.25 mg per kilogram of body weight plus an oral placebo, or an intravenous placebo plus oral aspirin at 300 mg [s1]. The double-dummy design meant neither patients nor staff could tell the groups apart.

The primary endpoint was vision recovery, defined as a best corrected visual acuity in the affected eye at 30 days of up to 0.7 logMAR — a threshold equivalent to 20/100 or better [s1]. Key secondary endpoints included a stricter acuity threshold of 0.5 logMAR (equivalent to 20/63 or better), the mean improvement in acuity, and a perimetry score measuring the field of vision [s1]. Safety endpoints centred on the recognised hazards of a clot-buster: symptomatic bleeding inside the skull, major bleeding elsewhere, and death [s1]. The trial is registered as NCT04526951 [s2].

What it showed

A total of 78 patients at 16 sites in six countries underwent randomisation, with 40 assigned to tenecteplase and 38 to aspirin [s1]. At 30 days, 8 patients (20%) in the tenecteplase group and 9 patients (24%) in the aspirin group had recovered vision by the primary definition [s1]. The difference ran slightly against the active drug: a risk difference of −3.7 percentage points, with a 95% confidence interval stretching from −22.0 to 14.7 and a p-value of 0.69 [s1]. That interval is wide, spanning a meaningful benefit and a meaningful harm alike, but it is centred close to zero — the signature of a treatment that did not move the needle. The secondary visual endpoints told the same story, with no substantial difference between the groups [s1].

The safety side did separate the arms, and not in tenecteplase's favour. There was a greater incidence of adverse events in the tenecteplase group, including one fatal intracranial haemorrhage — a bleed into the brain [s1]. In a condition where the drug did not demonstrably help, that is a consequential cost.

The limits

The most important limit is size. With 78 patients, TenCRAOS was small, and its confidence interval is correspondingly wide — it cannot rule out a modest benefit that a much larger trial might detect. But a small trial that finds no signal, and finds real harm, is not the same as an encouraging one. The result sits alongside earlier disappointment with thrombolysis for retinal artery occlusion and pushes in the same direction: the analogy to brain stroke, however intuitive, has not delivered. Vision recovery here was measured at 30 days, so these data speak to short-term outcome rather than the eventual settling point.

What to watch

TenCRAOS was an academically sponsored trial, led by Oslo University Hospital rather than a drug manufacturer [s1][s2] — the kind of independent evaluation that most often puts a hopeful idea to a fair test. Its conclusion is stated plainly by the investigators: intravenous tenecteplase given within 4.5 hours of onset did not result in significantly greater vision recovery at 30 days than oral aspirin, and was associated with serious safety concerns [s1]. For patients facing sudden monocular vision loss, the practical message is caution about clot-busting drugs offered outside a trial. Whether any thrombolytic, at any dose or window, can rescue the retina remains unproven; for now, the eye stroke has not yielded to the treatment that works for its cerebral cousin.

Sources

• [s1] A Randomized Trial of Tenecteplase in Acute Central Retinal Artery Occlusion. New England Journal of Medicine. 28 January 2026. doi:10.1056/NEJMoa2508515. PMID 41604638.
• [s2] ClinicalTrials.gov. TENecteplase in Central Retinal Artery Occlusion Study (TenCRAOS), NCT04526951. U.S. National Library of Medicine.