Merck Remigromig: Non-Inferior in DME, Higher PDR Rates

On September 24 Merck announced that remigromig, its Wnt-pathway agonist for diabetic macular edema, met the primary endpoint of the registrational Phase 2b/3 BRUNELLO trial. The press release called it the first novel mechanism of action for retinal vascular disease in more than two decades. Then, buried in the same release, came the caveat: higher rates of proliferative diabetic retinopathy, vitreous hemorrhage, and treatment discontinuations in the remigromig arms versus ranibizumab. The stock closed flat at $148.08. That is the correct reaction, and it tells you exactly how to think about this one.

The readout is thinner than the headline. BRUNELLO randomized 984 adults with diabetic macular edema 1:1:1 to 0.5 mg remigromig, 0.8 mg remigromig, or 0.5 mg ranibizumab, the standard Lucentis dose, every four weeks for a year. The primary endpoint was non-inferiority to ranibizumab on mean change in best-corrected visual acuity (BCVA) from baseline to week 52, scored on the ETDRS letter scale. Both remigromig doses cleared that bar. That is the good news, and it is nearly all Merck disclosed. No letter gains, no between-group differences, no non-inferiority margin, and no p-values. Merck and EyeBio advanced remigromig into BRUNELLO on the strength of the open-label Phase 1/2 AMARONE study, when the drug was still called Restoret, and this is the first randomized evidence that the Wnt approach holds up at all. The full dataset lands at AAO in New Orleans on October 10.

Here is the analytical problem with the headline. Non-inferiority to ranibizumab is a low bar. Lucentis is the oldest and most commoditized anti-VEGF on the market, and it has been losing share for years to Regeneron’s Eylea/aflibercept and, more recently, Roche’s Vabysmo (faricimab), a bispecific against VEGF-A and Ang-2 that is now the fastest-growing drug in the space and tracking toward roughly $7.7 billion by 2030. Matching ranibizumab says almost nothing about whether remigromig can compete with the agents physicians actually prescribe today. A drug that merely ties the weakest incumbent is a participation trophy, not a commercial threat.

Set aside the comparison for a moment and consider what remigromig actually does. Anti-VEGF drugs stop leaky blood vessels by blocking the growth signal that makes them leaky in the first place. Remigromig takes the opposite approach: it agonizes the Wnt pathway to repair and maintain the blood-retinal barrier itself. That is a genuinely different mechanism, and it is why the idea is interesting for the 40% of patients whose edema does not resolve on anti-VEGF alone. The practical question is treatment burden. Vabysmo’s biggest commercial advantage is its every-eight-to-sixteen-week dosing interval, which is why it has been converting patients off monthly Eylea and Lucentis. Remigromig was dosed every four weeks in BRUNELLO, and Merck has given no signal it will offer a longer interval. A monthly Wnt agonist that is merely non-inferior to monthly ranibizumab does not solve the convenience problem that made Vabysmo a multi-billion-dollar drug.

Then there is the safety signal, and it is the real story. The remigromig arms showed higher rates of proliferative diabetic retinopathy (PDR), vitreous hemorrhage, and adverse-event discontinuations. Merck and its investigators argue this is expected, because remigromig does not hit the VEGF pathway. Margaret Chang, a retina specialist and BRUNELLO investigator, told Modern Retina the finding is “likely a manifestation of the natural progression of diabetic retinopathy over time rather than an effect of the drug itself.” She may be right. But it cuts both ways. A drug that does not suppress VEGF also does not slow the ischemic drive that pushes DME patients toward PDR. Swap a patient off anti-VEGF and onto a Wnt agonist, and you are betting the Wnt pathway can carry the job alone. BRUNELLO is the first real test of that bet, and it is flashing yellow.

Why does Merck care at all? Because of a $3 billion decision made in May 2024, when Merck bought EyeBio for $1.3 billion upfront plus up to $1.7 billion in milestones, with remigromig as the lead asset. The DME market is real but not enormous: roughly $4.2 billion in 2025, with the leading agents Vabysmo and Eylea HD projected to generate a combined $13.2 billion by 2030 across their retinal indications. The prize Merck is actually chasing is the up to 40% of DME patients who do not fully respond to anti-VEGF and keep losing vision. That incomplete-responder niche is where a novel mechanism earns its keep, as an adjunct or an alternative. The trouble is Merck has not shown remigromig works better in those patients. It has only shown the drug is not worse than ranibizumab in the general DME population. That validates the mechanism; it does not demonstrate a commercial advantage.

This is where the mega-cap immateriality math takes over. Merck is a roughly $366 billion company at $148.08 a share, and even a $2 billion peak-sales scenario for remigromig is barely half a percent of the revenue base. The stock will not move on this drug, and it should not. The ophthalmology push matters to Merck as a diversification story, one more leg in the post-Keytruda rebuild covered in the MRK Q2 earnings deep dive, but it is a rounding error next to the oncology franchise that still dominates the pipeline, from the I-DXd B7-H3 ADC PDUFA to Welireg’s RCC expansion. Merck is also running a second, stronger eye program in MK-8748 (Tiespectus), a bispecific Tie2 agonist/VEGF inhibitor now in four registrational trials across wet AMD and DME. Tiespectus keeps the anti-VEGF mechanism remigromig lacks, which is why it is the more credible long-term play. The broader retina catalyst map lays out the full field.

This is the recurring pattern with Merck. The company is spending its Keytruda cash flow on a series of bolt-on bets, EyeBio among them, each too small to move the stock on its own but collectively an attempt to build a post-Keytruda growth stack before the oncology patent cliff hits. Remigromig’s BRUNELLO readout is a single data point in that broader effort, and it is more useful as a signal of whether Merck’s ophthalmology team can execute than as a reason to buy or sell the stock.

So what actually moves here? October 10. That is when Donald D’Amico of Weill Cornell presents the year-one data at AAO, and when investors finally see the letter gains, the p-values, the non-inferiority margin, and the event counts behind the safety signal. The binary is clean. If the safety data shows the PDR and vitreous hemorrhage rates are mild, reversible, and concentrated in patients who already had severe retinopathy at baseline, the EyeBio thesis survives and remigromig becomes a genuine adjunct for incomplete responders. If the signal is dose-limiting, or if efficacy is merely non-inferior with no hint of an edge in the harder-to-treat patients, remigromig is a niche product and Merck quietly prioritizes Tiespectus ahead of it. A second registrational DME trial, BAROLO, is already running, so this is not a one-shot readout either way.

The verdict: hold Merck and treat remigromig as a pipeline tell, not a stock catalyst. The honest probability is that remigromig gets approved, because non-inferiority on a BCVA endpoint with an acceptable safety profile is generally approvable. But approval of a non-inferior me-too in a market where Eylea and Vabysmo set the standard is not a commercial event. Do not own Merck for remigromig, and do not trade the AAO readout like a binary for the stock. The investors who should care are the ones who already own Merck for Keytruda and the oncology pipeline; for them, remigromig is a free option that either pays off in a decade or does not, and either way the stock barely notices.