New kind of blood thinner approved to help prevent repeat strokes

The US Food and Drug Administration approved Sunharbo (asundexian) on 9 November 2026 to lower the risk of another stroke in adults who have just had a mild-to-moderate ischaemic stroke or a high-risk transient ischaemic attack (TIA). Made by Bayer, it is the first approved medicine in a new class of blood thinners called factor XIa inhibitors. People take one 50 mg tablet a day alongside their usual antiplatelet medicine, such as aspirin or clopidogrel.

The approval rests on OCEANIC-STROKE, a trial of 12,327 people. Over a median of about 14 months of treatment, 6.2% of those on asundexian had an ischaemic stroke, against 8.4% on placebo. That is a 26% relative reduction. Major bleeding was similar across the whole trial, but it was more common in people over 75 and in the first 90 days, and bleeding from the gut was more frequent.

The decision comes after two high-profile setbacks for this drug class. Asundexian itself did worse than an existing blood thinner in people with atrial fibrillation, in a trial that was stopped early. And in August 2026 a rival factor XIa drug, milvexian, failed to prevent repeat heart attacks. The FDA's approval covers one specific use, and the label says plainly that Sunharbo is not approved for stroke prevention in atrial fibrillation.

What is factor XIa, and why target it?

Blood clotting relies on a chain of proteins, each switching on the next. Factor XI sits in a part of that chain that seems to matter more for clots forming inside blood vessels than for sealing a cut. People born with low levels of factor XI have fewer ischaemic strokes, and they bleed less than you might expect, the OCEANIC-STROKE investigators note in the New England Journal of Medicine.

Bayer's prescribing information says that blocking the activated form, factor XIa, reduces the amplification of thrombin, a key clotting enzyme. In animal models this hit harmful clotting harder than normal clotting. The hope is a blood thinner that prevents strokes with less of the bleeding that limits drugs such as warfarin, apixaban and rivaroxaban.

Hylton Joffe, who directs the FDA office that reviewed the drug, called the approval "a meaningful advance for patients and health care providers seeking a new option for secondary stroke prevention." Ashkan Shoamanesh of the Population Health Research Institute at McMaster University in Canada, a co-principal investigator of the trial, said in Bayer's statement that the risk of another stroke "is highest in the days and weeks after the initial event, and that risk persists over time."

Who is the approval for?

The FDA's announcement describes the patients as adults after a recent non-cardioembolic ischaemic stroke, meaning a stroke caused by a clot that did not come from the heart. The stroke must be mild to moderate, with a score of 15 or less on the NIH Stroke Scale, the checklist doctors use to rate stroke severity. People who had a high-risk TIA, sometimes called a mini-stroke, are also covered.

Sunharbo is meant to be added to antiplatelet treatment. It does not replace it. In the trial, people started within 72 hours of their symptoms and took asundexian or placebo on top of single or dual antiplatelet therapy chosen by their doctors.

The trial also required a sign that the stroke was linked to diseased arteries. According to the NEJM abstract, every patient had at least one of three features: a non-lacunar infarct on a brain scan (a stroke outside the small deep vessels), a history of atherosclerosis, or plaque visible on imaging of the brain or neck arteries. The average age was 68.

Strokes caused by atrial fibrillation are a different problem. Clots form in the heart, and the usual treatment is a full anticoagulant. The new approval does not cover them.

What did OCEANIC-STROKE show?

The main measure was the time to a first ischaemic stroke. These are the results reported by Bayer, the FDA and the NEJM paper.

The NEJM paper gives major bleeding as 1.9% against 1.7% of patients, with a hazard ratio of 1.10 (95% CI, 0.85 to 1.44). Bayer says the reductions in all strokes and in the combined cardiovascular outcome were mainly driven by fewer ischaemic strokes.

By our own arithmetic, the difference in ischaemic stroke is 2.2 percentage points. That means about 45 people would need to take asundexian for roughly 14 months to prevent one ischaemic stroke. Our guide to absolute and relative risk explains why the 26% figure and the 2.2 point figure describe the same result.

At the American Heart Association meeting in Chicago this month, Mukul Sharma of McMaster University and colleagues reported that the benefit and the bleeding rates looked similar across low, medium and high levels of baseline cardiovascular risk, according to the American College of Cardiology.

Asundexian had already stumbled once in stroke

OCEANIC-STROKE was not the drug's first test in stroke patients. In the phase 2b PACIFIC-Stroke trial, published in The Lancet in 2022, 1,808 people with an acute non-cardioembolic stroke took asundexian at 10, 20 or 50 mg, or placebo, on top of antiplatelet treatment. The main outcome combined new silent brain infarcts seen on MRI with repeat strokes at 26 weeks. It was almost identical across the groups: 19% on placebo and 20% on the 50 mg dose.

That trial was designed to find a dose, and its main outcome was dominated by small silent infarcts on scans. The phase 3 trial that followed used the 50 mg dose, counted only symptomatic ischaemic strokes, enrolled nearly seven times as many people (12,327 against 1,808), and limited entry to patients with signs of artery disease or a non-lacunar infarct. With those changes, the drug showed a clear benefit.

What are the side effects and cautions?

Bleeding is the main concern with any blood thinner, and the label gives more detail than the headline figure. In Bayer's summary of the prescribing information, major bleeding rates were:

• 4.5 against 3.1 per 100 patient-years in the first 90 days of treatment
• 2.6 against 1.7 per 100 patient-years in people older than 75
• 1.9 against 1.5 per 100 patient-years in people planned to take dual antiplatelet therapy
• 2.9 against 1.0 per 100 patient-years in the high-risk TIA subgroup

Bleeding from the gut was more common on asundexian. Bayer says 74% of those bleeds were classed as minor or clinically relevant non-major, and the most common site was the rectum.

Diarrhoea was the most frequent side effect, in 3.4% on asundexian and 2.8% on placebo. The drug can raise liver enzymes. Most of these rises were mild, and serious liver injury with jaundice was reported in fewer than 0.02% of patients. Adverse events of any kind affected 69.3% of people on asundexian and 70.1% on placebo, and serious adverse events 19.2% and 19.5%.

The FDA lists several practical cautions. Sunharbo is not recommended in people with moderate or severe liver impairment. It should not be combined with medicines that strongly speed up its breakdown, such as carbamazepine, because they can make it less effective. It can raise levels of digoxin. Doctors should stop it at least 24 hours before planned surgery, and hold it for at least 24 hours after clot-busting treatment for an acute stroke. Breastfeeding is not recommended during treatment.

Why did asundexian fail in atrial fibrillation?

In OCEANIC-AF, Bayer compared asundexian 50 mg once daily with apixaban in 14,810 people with atrial fibrillation at high risk of stroke. An independent monitoring committee recommended stopping the trial early, in November 2023. Stroke or systemic embolism occurred in 1.3% of the asundexian group and 0.4% of the apixaban group (hazard ratio 3.79). Major bleeding was much lower with asundexian, at 0.2% against 0.7%, but the drug clearly did not protect as well as the standard treatment.

The two trials asked different questions. In atrial fibrillation, asundexian had to match apixaban, one of the most effective drugs for preventing clots that form in the heart. In OCEANIC-STROKE, it was added to antiplatelet medicine and compared with placebo, in strokes caused by diseased arteries. A drug can fall short of a strong anticoagulant in one condition and still add benefit on top of aspirin-type drugs in another. That is our reading of the trial designs, not a conclusion the investigators have published.

It does mean the approval should be read narrowly. Asundexian has shown benefit in one group of stroke patients. It has shown harm, in the form of more strokes, when used in place of apixaban in atrial fibrillation.

How does this fit with milvexian's failure?

Milvexian, developed by Bristol Myers Squibb and Johnson & Johnson, also blocks factor XIa, but it is a different molecule given twice daily. In LIBREXIA-ACS it did not reduce heart attacks, strokes or cardiovascular deaths in people who had just had a heart attack or unstable angina, as we reported from ESC Congress 2026.

Those results do not cancel each other out. The trials tested different drugs, at different doses, in different diseases. Heart attacks are driven mostly by platelets clumping on ruptured plaque in coronary arteries, while OCEANIC-STROKE focused on strokes from diseased brain and neck arteries. Milvexian's own stroke trial, LIBREXIA-STROKE, had not reported results when we checked on 10 November 2026.

The lesson so far is that factor XIa drugs need to be judged one use at a time. A success in secondary stroke prevention is not evidence that the class works in heart attacks or atrial fibrillation.

Can people in the UK or Europe get asundexian?

Not yet. Bayer says asundexian is under review by the European Medicines Agency and by Japan's health ministry. China approved it on 26 October 2026, according to Bayer's announcement at the time. Bayer has not announced a UK application to the MHRA, and NICE has not appraised it. Our explainer on how new medicines get approved covers each of those steps.

For now, preventing another stroke in the UK relies on established treatments. These include antiplatelet medicines, statins, stopping smoking and controlling blood pressure. On that last point, the recent OPTIMAL-STROKE trial found that a lower blood pressure target cut further heart attacks and strokes in stroke survivors.

In the US, Bayer expects Sunharbo to start shipping in mid-to-late November and says patients who cannot afford it may qualify for its patient assistance foundation.

Questions to ask your stroke team

If you have had a recent ischaemic stroke or TIA, these questions may help:

• Was my stroke caused by a clot from the heart, or by disease in my arteries?
• Am I on single or dual antiplatelet therapy, and for how long?
• Would a factor XIa drug be suitable for me if it becomes available here?
• What is my risk of bleeding, given my age and other medicines?
• What else can I do to lower my risk of another stroke?

This article is for general information and is not medical advice. Do not start, stop or change any blood thinner without talking to your doctor.

What remains uncertain?

Sunharbo is the first new kind of blood thinner for stroke prevention in years, and OCEANIC-STROKE gives clear evidence that it lowers the risk of another ischaemic stroke in the patients it studied. The open questions are about balance and reach. Bleeding was higher early on and in older people. The trial lasted a little over a year on average, and the atrial fibrillation result shows how much the setting matters.

The next markers are the European and Japanese decisions, and the results of milvexian's stroke trial. Together they should show whether factor XIa inhibition earns a lasting place in stroke care.