Shares of EyePoint Pharmaceuticals plunged about 70% after its lead eye drug missed the main goal of a pivotal trial, erasing nearly $1 billion in market value.

Durayvu, an intraocular injection for wet age-related macular degeneration, failed to demonstrate non-inferiority to aflibercept on best-corrected visual acuity across the roughly 400-patient LUGANO study — the first of two pivotal Phase 3 trials.

The disease and its treatment burden

Wet AMD is a leading cause of vision loss in older adults. Abnormal blood vessels grow beneath the retina and leak fluid and blood, damaging the macula responsible for central vision — reading, faces, driving.

Anti-VEGF drugs injected into the eye stop that vessel growth and have transformed outcomes from near-certain progressive blindness to preserved vision for many patients.

The cost is the regimen. Injections directly into the eye, repeated every one to two months, indefinitely. Patients are typically elderly and require transport to a specialist each time, and the cumulative burden causes many to fall behind on treatment — at which point vision is lost and does not return.

That is why long-acting options are commercially attractive: the unmet need is not efficacy but frequency.

What non-inferiority means, and why missing it is decisive

A non-inferiority trial does not attempt to show a new drug is better. It attempts to show it is not meaningfully worse than the standard, on the reasoning that equivalent vision with fewer injections is a win.

That is a lower bar than superiority, which is precisely why missing it is so damaging. A drug failing to demonstrate equivalence against an established comparator has failed the easier test, and there is no consolation framing available.

The company’s explanation

EyePoint attributed the miss to an “asymmetric cohort” of nine Durayvu-treated patients who experienced unexplained vision loss unrelated to wet AMD, and said a post-hoc analysis excluding that group would have met non-inferiority.

It also highlighted durability: about three-quarters of patients went roughly eight months without supplemental anti-VEGF injections. CEO Jay Duker said the results still present “a compelling case,” with a potential FDA filing in the first half of 2027 pending the second trial, LUCIA.

Why post-hoc exclusion carries little weight

The argument is not absurd, and it is close to unusable as evidence.

Removing patients from an analysis after seeing the results is the definition of a post-hoc analysis, and it is unreliable because the removal criteria are chosen with knowledge of which patients caused the problem. Almost any trial can be made to succeed by excluding its worst outcomes.

There is a more specific concern. Nine patients in the treatment arm experiencing unexplained vision loss is itself a finding that requires explanation. Describing it as unrelated to wet AMD does not establish it was unrelated to the drug — and for an intraocular injection, unexplained vision loss in treated patients is exactly the safety signal that would matter most.

The company may be right that these were coincidental. The trial cannot demonstrate it.

The durability data is the real asset

Three-quarters of patients going roughly eight months without supplemental injections is a genuinely strong result, and it addresses the actual unmet need.

Against a standard requiring injections every one to two months, eight-month durability would transform the treatment experience — if the vision outcomes hold. The problem is that durability is worthless without demonstrated efficacy, since a long-acting drug that preserves vision less well is not an improvement.

What happens next

Analysts at RBC Capital Markets said the LUCIA results “will be needed” for clarity, warning that a second failure would be “far more problematic.”

That is the position precisely. If LUCIA meets non-inferiority cleanly, the LUGANO miss becomes arguable as an anomaly driven by an unlucky cluster, and the durability data carries the filing.

Why long-acting eye drugs keep coming up short

EyePoint is not the first company to find that extending dosing intervals in wet AMD is harder than it appears, and the reason is a genuine pharmacological tension.

Anti-VEGF therapy works by keeping drug levels above a threshold that suppresses abnormal vessel growth. Extending the interval means either delivering more drug initially, which raises tolerability and safety questions in a small enclosed organ, or releasing it more slowly, which risks levels falling below threshold late in the cycle.

Vision lost during a period of inadequate suppression does not fully return, so the cost of dipping below threshold is asymmetric — unlike most chronic conditions, where a brief lapse in control is recoverable.

That asymmetry is why non-inferiority trials in this setting are unforgiving. A drug can perform well for most of its interval and still lose to a comparator dosed more frequently, because the comparator never has a weak window. It also explains why durability data alone, however impressive, cannot carry a filing.

If LUCIA also misses, the post-hoc explanation collapses entirely, and a 70% share decline will have been an underreaction. Investors reacted to the endpoint miss, not the post-hoc arithmetic — which is the correct way to read a trial. Business news, not investment advice.