Kodiak Sciences reported on Sept. 28 that its retina drug Zenkuda matched aflibercept on vision in the Phase 3 DAYBREAK trial and held 54 percent of wet age-related macular degeneration patients on a six-month dosing interval through the first year. The same readout produced a second positive result, for a bispecific antibody called tabirafusp-ted. Kodiak's shares more than doubled, adding more than $3 billion in market value, while Regeneron fell 4 percent.
The result is real, and the market is treating it as settled. The more useful question is narrower than the one the share price answered. This molecule has been through Phase 3 twice before, and the history says which figure in the release carries the most weight.
What DAYBREAK found
DAYBREAK randomized 690 patients with treatment-naive wet AMD across three arms: Zenkuda, tabirafusp-ted, and aflibercept. The trial registry lists the primary endpoint as change in best-corrected visual acuity at week 48, tested for non-inferiority rather than superiority.
Both experimental drugs met it. Kodiak reported a p-value of 0.0007 for Zenkuda against aflibercept, 0.0036 for tabirafusp-ted on vision, and less than 0.0001 for the bispecific on a key anatomical secondary endpoint. The release does not disclose the mean letter gains for either arm. CTOL Digital reported Zenkuda at 7.2 letters against aflibercept's 7.6, a small numerical gap that a non-inferiority margin is built to treat as immaterial.
The durability figure is the one the company is selling. After four monthly loading doses, Zenkuda patients moved to an individualized schedule running from every four weeks to every 24, with retreatment triggered by any fluid visible on OCT imaging. On that protocol, 54 percent reached a six-month interval by the end of year one.
The safety numbers came in cleaner than the comparator's. Kodiak reported no intraocular inflammation in the Zenkuda arm and a cataract adverse event rate of 0.5 percent, against 0.9 percent for aflibercept. Tabirafusp-ted came in at 0.4 percent inflammation and no cataracts.
The program was wound down once already
Zenkuda is tarcocimab tedromer, and it has been here before.
In February 2022, Kodiak reported that the Phase 2b/3 DAZZLE study in wet AMD missed its primary endpoint, failing to show non-inferiority to Eylea in a 559-patient trial. Intraocular inflammation ran 3.2 percent on tarcocimab against zero on aflibercept, and the shares fell about 80 percent.
The second failure is the instructive one. In July 2023, the Phase 3 GLEAM and GLIMMER studies in diabetic macular edema missed their primary efficacy endpoints, and the company wound down the tarcocimab program. Kodiak's own filing describes a drug that did not fail to last. It lasted, and it lost on something else. Development resumed that November after positive results in the BEACON study in retinal vein occlusion and the GLOW1 study in diabetic retinopathy.
The number that matters is the cataract rate
The quarterly report Kodiak filed for the period ended Sept. 30, 2023 records what happened in GLEAM and GLIMMER in unusual detail. The two studies were identically designed and enrolled 460 and 457 treatment-naive patients with diabetic macular edema. Half of the tarcocimab patients were on every 24-week dosing at the primary endpoint, two-thirds reached at least one six-month interval, and three-quarters reached an interval of five months or longer.
Durability, in other words, was never the problem. Cataracts were: 19 percent on tarcocimab against 9 percent on aflibercept at the primary endpoint, an imbalance the company said contributed meaningfully to the visual acuity decline that sank both studies.
That is why the 0.5 percent cataract rate in DAYBREAK matters more than the 54 percent. Kodiak had already shown it could stretch an injection interval in a pivotal trial. What it had not shown, until this readout, was that it could do so in wet AMD without the lens opacity that erased the vision benefit the last time. The same filing notes that DAYLIGHT, the company's other wet AMD study, showed no cataract imbalance despite monthly dosing, which points to something specific about the diabetic macular edema population or the dosing density there rather than to the molecule itself.
The company's chief medical officer, J. Pablo Velazquez-Martin, framed the readout as the end of a long road. "Four years and seven months after our first Phase 3 readout," he said, the profile has come into focus through disciplined learning.
What 54 percent measures
The durability comparison in DAYBREAK is not like for like, and the protocol is why.
Zenkuda patients were on an as-needed schedule governed by an imaging algorithm that retreated for any detectable fluid. Aflibercept patients were dosed per label, which for the 2 mg formulation means a fixed eight-week interval. The trial compared a regimen free to stretch to six months against one that could not stretch at all. The Zenkuda arm had the option to look durable. The aflibercept arm did not.
Charles Wykoff, a retina specialist at Houston Methodist named in the release, said the safety profile appeared consistent with expectations for aflibercept. David M. Brown of Retina Consultants of America, an investigator on the trial, said more than half of patients were maintained on 24-week dosing through year one. Both are quoted in the company's own announcement.
Two limits belong next to that number. Fifty-four percent is a majority but not a rule; the other 46 percent were on shorter intervals, and the release does not break out the distribution. And durability was measured at one year, in a disease treated for the rest of a patient's life.
The comparator is the dose the market left behind
The choice of comparator is the sharpest constraint on what the result establishes.
The DAYBREAK aflibercept arm received 2 mg every eight weeks. Regeneron's Eylea HD is aflibercept at 8 mg, approved for intervals of eight to 16 weeks, with 77 percent of patients in the PULSAR trial reaching 16-week intervals beyond the first year. Roche's Vabysmo blocks both VEGF-A and Ang-2 and is approved for four to 16 weeks, with roughly 45 percent of patients reaching 16-week dosing in its pivotal trials. Real-world data on patients switched to Vabysmo shows intervals lengthening from four weeks to an average of 9.4.
Zenkuda's six-month dosing is longer than either competitor's approved maximum, which is the strongest argument for it. No trial has run it against Eylea HD or Vabysmo, and Kodiak does not claim one. The company's counter is that trial-defined intervals overstate what the newer drugs deliver when patients miss visits, and that a durability advantage measured under stricter retreatment criteria would look better, not worse, outside a protocol. That may hold. It is the kind of claim a head-to-head study settles.
What the filing will decide
Kodiak plans to submit a three-indication application in the fourth quarter, covering wet AMD, diabetic retinopathy, and retinal vein occlusion, built on five positive Phase 3 studies: DAYBREAK and DAYLIGHT in wet AMD, GLOW1 and GLOW2 in diabetic retinopathy, and BEACON in retinal vein occlusion. On that schedule, approvals would land in late 2027 or early 2028. Kodiak owns full rights to all three of its late-stage assets and has no partner.
The economics of a durability drug are less favorable than the clinical case. Retinal injections run roughly $1,800 to $2,500 per dose, so a drug given twice a year instead of six times needs a substantially higher price per injection to produce the same revenue per patient. Regeneron and Roche have reason to defend intervals that already reach 16 weeks, and payers will want the comparison that has not been run.
Where durability does pay, and where the argument is strongest, is adherence. Real-world retina patients miss visits, and a regimen built around six-month intervals forgives a missed appointment in a way that one built around eight weeks does not. That is a benefit Kodiak can support without a head-to-head trial. Tabirafusp-ted is being tested for superiority rather than non-inferiority in a roughly 910-patient Phase 3 study in diabetic macular edema, and a third asset, KSI-101, reports topline data in December.
What would settle the rest is a study against Eylea HD or Vabysmo, or several years of real-world data after a launch. Neither exists yet. Kodiak will file first, and the retina specialists who have patients stable on Vabysmo will decide whether a non-inferiority win against an older dose, in a molecule that failed twice, is enough to move them.
Primary sources
- Kodiak Sciences, Zenkuda and tabirafusp-ted Meet Primary Endpoints in Pivotal DAYBREAK Trial in wAMD, filed as Exhibit 99.1 to a Form 8-K on Sept. 28, 2026, for the trial design, the p-values, the durability and safety figures, the platform description, the pipeline status, and all company and investigator quotations.
- ClinicalTrials.gov, A Study to Evaluate the Efficacy and Safety of Tarcocimab Tedromer and Tabirafusp Tedromer Compared to Aflibercept in Participants With Neovascular (Wet) Age-related Macular Degeneration (wAMD) - DAYBREAK, NCT06556368, for the three-arm design, the aflibercept 2 mg comparator, the planned enrollment, and the week 48 primary endpoint.
- Kodiak Sciences, Quarterly Report on Form 10-Q for the quarterly period ended September 30, 2023, for the GLEAM and GLIMMER enrollment, the 24-week dosing rates, the cataract imbalance, the wind-down of the tarcocimab program, and the absence of a cataract signal in DAYLIGHT.
- FierceBiotech, Kodiak's Eylea rival hits 'major disappointment' with failure in first of 6 phase 3 studies, Feb. 23, 2022, for the DAZZLE enrollment, the missed endpoint, and the inflammation imbalance.
- BioSpace, Kodiak Sciences stock soars 166% after 2 assets match Eylea in Phase 3 trial, Sept. 28, 2026, for the share price move and the market reaction.
- Healio, Two investigational agents for wet AMD meet phase 3 primary endpoints, Sept. 28, 2026, for the readout summary and the three-indication filing plan.
- Inside Precision Medicine, Phase III success pushes Kodiak's eye treatment towards approval, Sept. 28, 2026, for the DAYBREAK enrollment of 690 patients and the six-month dosing rate.
- CTOL Digital, Kodiak's DAYBREAK Win Beat an Outdated Eylea Dose, Sept. 28, 2026, for the mean letter gains, the Regeneron share decline, and the per-injection cost range.
- Medscape, Switching to Faricimab Offers Vision Gains, Fewer Injections, 2026, for the real-world Vabysmo dosing intervals.