A single injection of Connect Biopharma's antibody rademikibart, given to patients in the middle of an acute COPD exacerbation, cut the rate of treatment failure over the next 28 days by 81 percent compared with placebo, the company said on September 30. Seventy-eight patients received the drug and 81 received a placebo, all of them on top of standard care. Two treated patients met the study's failure definition against 11 in the placebo group, a difference with a p-value of 0.0122.
The result is the strongest the program has produced, and it comes with a complication three weeks in the making. The same 28-day treatment-failure endpoint had just missed in asthma, in a study that used the same design, the same dose and the same definition of failure. The company now has to walk into a meeting with the Food and Drug Administration carrying both.
The COPD topline release says Connect plans to engage the agency on a Phase 3 program. The asthma release from September 15 adds the detail that matters most for what follows: the two readouts "should help determine a potential Phase 3 endpoint and sample size requirements." After two Phase 2 studies in two diseases, the endpoint is still an open question.
The shot is given during the flare, not between flares
COPD affects nearly 16 million diagnosed adults in the United States, and it has exactly one approved biologic. Dupilumab is a maintenance therapy, 300 milligrams under the skin every two weeks, taken continuously by patients with an eosinophilic phenotype whose disease is not controlled on inhaled triple therapy. The Phase 3 BOREAS trial behind its 2024 approval enrolled 939 patients and reduced the annualized rate of moderate or severe exacerbations by about 30 percent over 52 weeks.
Rademikibart binds the alpha subunit of the interleukin-4 receptor, the shared piece of the receptors for IL-4 and IL-13, the two signals that drive type 2 inflammation. That is the same target dupilumab binds. The difference in the Seabreeze STAT program is not the molecule class but the moment of use. Patients were enrolled during an acute exacerbation, at an urgent outpatient visit, in an emergency department or in a hospital bed, and each received one 600-milligram dose under the skin alongside steroids, antibiotics and bronchodilators.
The logic rests on the window after a flare rather than the flare itself. Patients who go to an emergency department for a COPD exacerbation are at their most vulnerable in the weeks that follow, with roughly one in five readmitted within 30 days. A single dose that stabilizes that period would be a different product from a maintenance injection, aimed at a different moment in the disease.
Treatment failure is a composite, and every part of it favored the drug
The primary endpoint counted four events inside 28 days: death from any cause, admission or readmission to a hospital for COPD, a visit to an emergency department or an unscheduled medical visit for worsening symptoms, and the need to intensify drug treatment. A patient met the endpoint on any one of them, and the investor presentation breaks the count apart.
Eleven of the 81 placebo patients failed, or 13.6 percent. Two of the 78 treated patients failed, or 2.6 percent. The components moved in one direction: four placebo patients were admitted or readmitted to a hospital and no treated patient was, three made an emergency department visit for worsening symptoms and none did, two in each arm had an unscheduled visit and nothing more, and two placebo patients needed intensified drug treatment without any visit.
The two failures in the rademikibart arm were both that mildest category, an unscheduled visit with no admission behind it. Counting patients who returned to a hospital or an emergency department, the split was seven against zero, with a p-value of 0.0137. On new moderate-to-severe exacerbations through Week 4, the count was seven against one, an 85 percent reduction with a p-value of 0.0301.
One number in the presentation is more conservative than the headline and belongs beside it. Counting failure events rather than patients, the reduction was 77 percent, with a p-value of 0.0373. The share of patients who failed fell further than the raw event count did, which happens when a few placebo patients accumulate more than one failure each.
Zero events, 166 screen failures, and a map that leaned east
The 100 percent reduction in hospital and emergency department returns rests on seven events in one arm and none in the other. With 78 patients in the treatment arm, one readmission would have turned that figure into 1.3 percent and changed the sentence everyone quotes. The company's own note says the study was powered for treatment failure and not for the others, so the exacerbation and return-visit figures describe what happened in this cohort rather than a test the trial was built to pass.
Safety ran in the drug's favor. Treatment-emergent adverse events were reported in 14.1 percent of treated patients against 25.9 percent of placebo patients, serious adverse events in 2.6 percent against 8.6 percent, and no patient stopped treatment because of a side effect. One patient in the rademikibart arm died, on Day 37, which the company reported as unrelated to treatment and described as a participant with a history of ventricular arrhythmias who woke during the night and fell. A single death in a 159-patient study carries little statistical weight in either direction, and a rare risk would surface in the larger safety database a Phase 3 would generate.
The trial screened 325 patients and randomized 159. Of the 166 who did not get in, 99 failed the eligibility criteria and 63 had no exacerbation while they were being screened, a reminder of how narrow the entry conditions were: patients had to be in an acute flare at enrollment and had to have a blood eosinophil count of at least 300 cells per microliter.
Among the 159 who enrolled, 82 were treated in Serbia, 49 in Georgia, 18 in the United States, seven in Argentina and three in Australia. The mean age was about 65, and roughly 45 percent were current smokers. A Phase 3 meant to support a United States label will need enough domestic enrollment to show the effect in the population that would use the drug, and the geography of this study is one of the things the FDA meeting will cover.
Three weeks earlier, the same endpoint missed and lung function moved the other way
The asthma study enrolled 160 patients with the same entry requirements and the same endpoint. Rademikibart reduced treatment failure by 66 percent, and the result fell short of significance, with a p-value of 0.153. Connect attributed the miss to a lower rate of treatment failure than projected in both arms, which is the statistical version of saying the control group did better than expected.
The secondary endpoint in asthma went the other way, and strongly. Post-bronchodilator FEV1, the volume of air a patient can force out in one second, improved by 250 milliliters on the drug against 120 milliliters on placebo at Day 7, a 130-milliliter difference with a p-value of 0.023. Connect has said it proposes to use that lung-function measure as the primary endpoint in a Phase 3 asthma program.
In the COPD study the pattern inverted. The composite endpoint hit and lung function did not. The treated group improved FEV1 by about 70 milliliters more than placebo at Week 4, which the release calls clinically meaningful and which came with a p-value of 0.1607, well short of significance. The Day 7 lung-function measure that carried the asthma trial was not significant in COPD either.
Set the two studies side by side and the endpoint that failed in asthma is the one the company now wants to build on, while the endpoint that succeeded in COPD is not the one anyone proposes to use in asthma. Whether the FDA accepts a program whose registrational measure differs by disease, and a single-dose acute treatment at all, is the substance of the meeting Connect is seeking.
The market opened at the high and sold the rest of the day
Connect shares closed at 99 cents on September 29. They opened at $1.89 on September 30 and closed at $1.10, up 10.9 percent, on 164 million shares traded against 1.6 million the day before. The gap between the opening print and the close is the market's own summary of the news: enough to reprice the company upward, not enough to hold the first number.
The constraint the data do not solve is financial. Connect reported $31.5 million in cash, equivalents and short-term investments as of June 30 and said in its August quarterly release that the balance would fund at least one year of operations. A Phase 3 program in two diseases, with a safety database large enough to satisfy the agency and enough United States sites to support a label, costs considerably more than that, and the company filed a shelf registration in May.
The cost case is doing as much work as the statistics
The presentation also makes an economic argument that is more ambitious than the clinical one. Using 2025 national claims data for patients aged 40 and over, it counts 3.46 million COPD-related emergency department visits a year, assigns 46 percent of them to type 2 inflammation, notes that 58.9 percent end in admission, and puts the 30-day return rate between 11 and 20 percent depending on how the first visit ended. It then applies the trial's 81 percent reduction to every one of those returns and arrives at $6 billion in savings if every such patient received the drug.
Two things about that number matter. The claims model defines type 2 inflammation as a blood eosinophil count above 150 or a documented atopic condition, while the trial enrolled patients at 300 and above, so the savings figure applies a result measured in a narrower group to a broader one the trial did not study. And the model assumes every eligible patient is identified and treated at the moment of the flare, which is an operational problem before it is a financial one. Connect began a 40-patient study in August to test an intravenous push version of the drug, the form an emergency department would want, and that study is the tell for how far the $6 billion sits from a hospital's purchasing decision.
What the COPD result establishes is narrower than the reaction to it suggests and more useful than the asthma result was. A single dose of a type 2 blocking antibody, given at the start of an exacerbation, reduced a composite of bad outcomes in a small randomized trial, with the benefit concentrated in the events patients and hospitals care most about. What it does not establish is which measurement a Phase 3 should be built around, how long the effect lasts past the first month, or whether an acute biologic can earn a label in a disease whose approved biologic is taken every two weeks for years. Those are the questions the FDA meeting exists to answer, and until it happens, the data support a meeting rather than a filing.