Hope Sterling: Juniper, tell me if this is dramatic — but I have been sitting with this image all week of someone in their late sixties, Thursday afternoon, neurologist's office, three days after their words just stopped working mid-sentence. And their hand went numb. And then it was fine. And the doctor calls it a TIA — a mini-stroke — and slides a prescription across the desk for something called Sunharbo.
Juniper Vale: That's not dramatic at all — that's literally the scenario this drug was built for.
Hope Sterling: And what I want to figure out today is — is that prescription actually good news? Like, real good news, not press-release good news?
Juniper Vale: That's the whole question. And I think you have to start with what this drug actually does, before you can answer whether it's good news for that specific person.
Hope Sterling: Okay — go. Plain language, please.
Juniper Vale: Think of it this way. Your blood has two different clotting systems — one protects you when you bleed, like a cut or a scrape, and one is involved in the kind of clot that forms inside your arteries and causes strokes. What Sunharbo is trying to do is block just that second, dangerous pathway — without shutting down the protective one. That's what makes it a new class. It's a factor XIa inhibitor, the first one the FDA has ever approved for this, and they said yes on September 11, 2026.
Hope Sterling: Wait — first ever approved in the whole U.S.?
Juniper Vale: First ever. And it's a once-daily pill — 50 milligrams — and it goes on top of the aspirin and clopidogrel that someone after a TIA is typically already taking. It doesn't replace them. It stacks.
Hope Sterling: Okay but — stacks on top. So this person's already on aspirin and clopidogrel, and now we're adding a third thing. And the number everyone's leading with is 26% reduction. That sounds huge.
Juniper Vale: It does. And it's real — the trial was OCEANIC-STROKE, 12,327 people, 37 countries, median 14 months of follow-up. Andrew Russman at Cleveland Clinic, one of the investigators, called it a likely paradigm shift. The results landed in the New England Journal of Medicine on April 16, 2026. That hazard ratio of 0.74 — that's where the 26% comes from, and it's not invented.
Hope Sterling: So what's the problem?
Juniper Vale: The problem is — I mean, it's not a problem exactly, it's more like — a physician on X, @DrDeepMD, looked at the same data and said: the number you actually need is the absolute difference. And that number is 2.2 percentage points.
Hope Sterling: Wait. So it went from like 8-ish percent of people having a stroke down to 6-ish percent?
Juniper Vale: Exactly that. Roughly 8.4 to 8.8 on placebo, dropping to about 6.2 to 6.6 on asundexian. So yes — 26% sounds like a sweeping win. But the lived version is: out of 100 patients who take this drug for over a year, about 2 fewer have a stroke.
Hope Sterling: Okay, I genuinely don't know if that's big or small. Like — is 2 people out of 100 a lot?
Juniper Vale: And that is — wait, that's actually the question that the press releases skipped. The FDA approval messaging, the Bayer announcements — they led with 26%. The International Stroke Conference presentation in February, same thing. Nobody put the 2.2 in the headline. And I'm not saying it's dishonest — both numbers come from the same data — but they create genuinely different conversations between a doctor and a patient.
Hope Sterling: Because 26% feels like — yes, obviously, give me this drug. And 2 out of 100 makes you actually stop and think.
Juniper Vale: And that pause — that's actually where the safety story comes in, because people land on 2 out of 100 and then say, okay, but at least it's safe. And the aggregate number kind of backs that up. Overall major bleeding was 1.9% on asundexian versus 1.7% on placebo. Hylton Joffe at the FDA — he directs the Office of Cardiology, Hematology, Endocrinology and Nephrology at CDER — he signed off on the approval, priority review, real unmet need recognized. And that 1.9 versus 1.7 is not statistically different.
Hope Sterling: The bleeding headline is basically — fine?
Juniper Vale: On the surface. But then you crack open the subgroups and — I mean, it stops looking quite as clean. Patients over 75, elevated major bleeding risk. The first 90 days specifically — that's the window where someone is on Sunharbo AND aspirin AND clopidogrel simultaneously. Triple antithrombotic load. That's when the signal appears.
Hope Sterling: Wait — so the people most likely to BE on this drug, older adults after a TIA, are the exact people where the bleeding data gets bumpier?
Juniper Vale: That's the footnote that @bmorecardiology flagged on X — specifically called out elderly subgroup signals as a reason not to treat the aggregate approval number as the whole story. And then there's the GI piece — gastrointestinal bleeding was more frequent in the asundexian arm. And honestly, the detail that stands out — diarrhea. Most common adverse reaction, hitting at least 3% of patients.
Hope Sterling: Three percent — like, picture a retirement community, a 75-year-old woman six weeks into the drug, and she's not calling her neurologist about diarrhea. She's just quietly stopping the pill.
Juniper Vale: Adherence, yes. And that's not in the trial numbers.
Hope Sterling: So we've got a benefit that's smaller than it sounds AND a safety story with a real asterisk for the population most likely to get prescribed it. That's — I don't know, that feels like the individualized risk-benefit thing isn't optional, it's the whole job.
Juniper Vale: And there's a layer under even that — because asundexian already failed in a totally different stroke population before this trial ever happened, and understanding why that failed versus why this worked tells you something important about where the paradigm-shift framing actually holds. We'll get there.
Hope Sterling: Wait — the same drug that already failed? Like, Bayer tried this before and it didn't work?
Juniper Vale: November 2023. Bayer discontinued a large asundexian trial in atrial fibrillation patients — the drug went up against a competitor and the results just weren't strong enough. They stopped it.
Hope Sterling: So the same molecule. Same pill, basically. And it flopped in AF patients, and now it's FDA-approved and everyone's calling it a paradigm shift?
Juniper Vale: And that's — I mean, that's actually the thing that tightens the Andrew Russman quote for me. Because his 'paradigm shift' language might be exactly right for this slice — non-cardioembolic ischemic stroke, high-risk TIA, that specific population. But the AF failure tells you something: FXIa inhibition isn't a universal clot-stopper. It didn't work where clots form in a fibrillating atrium. It did work where the mechanism of stroke is different. So either the biology is narrower than the original vision, or Bayer just found the one subtype where this pathway is load-bearing.
Hope Sterling: And Sunharbo is explicitly not approved for AF patients now, right? Like, that's just — off the label.
Juniper Vale: Completely off. The indication is locked to recent mild-to-moderate non-cardioembolic ischemic stroke or high-risk TIA. AF is not on there. And Bayer knew exactly what that meant — they needed a growth driver after the AF program collapsed, and this approval is real, but it is narrower than what they originally built toward.
Hope Sterling: That's — okay, that reframes it a lot. Like, picture a cardiologist who's been watching this drug for three years, sees the AF trial die, and then sees the OCEANIC-STROKE data come out in the New England Journal of Medicine and thinks — oh, so that's where it actually lives.
Juniper Vale: And then that cardiologist has to remember — when they're prescribing it — there's a drug-interaction screen built in. Combined P-glycoprotein and strong CYP3A4 inducers aren't recommended alongside Sunharbo. So that's one more thing the prescriber has to run before they write it. It's not complicated, but it's another layer on a drug that's already narrowly indicated.
Hope Sterling: So the paradigm shift has, like, a pretty specific zip code.
Juniper Vale: A very specific zip code. And the thing that sits with me — the FXIa idea is genuinely elegant. Target the amplification pathway, spare the protective one. That promise is real. But the data is telling us the decoupling isn't complete, especially past 75. The mechanism doesn't fully deliver on itself in the people most at risk.
Hope Sterling: So the patient who actually benefits most — it's probably someone younger, lower bleeding history, high recurrence risk. Not the average person walking out of a stroke clinic.
Juniper Vale: That's the honest read. And I think the next real question for stroke centers isn't whether Sunharbo works — it's whether anyone builds the infrastructure to ask 'for whom' before writing the prescription. Because if no one does that work, it just becomes a reflexive third agent.
Hope Sterling: And we're back to that person in the neurologist's office. Except now the question is different — it's not 'is there something new?' It's 'is this the right something for you specifically.'
Juniper Vale: Yeah. That's the whole job. Thanks for thinking through it with me.