On November 9, Dr. Hylton Joffe, who directs the FDA office that reviewed Bayer’s new blood thinner, called its approval “a meaningful advance for patients and health care providers.” The drug is asundexian, now sold as Sunharbo. A few paragraphs further down, the agency’s release carries a less festive sentence. Overall major bleeding was similar to placebo, the FDA wrote, but it “was increased with Sunharbo in patients older than 75 years and during the first 90 days of treatment.”

Those first 90 days are when a patient who has just had a stroke is most exposed to another one, and the pivotal trial tested them as a prespecified secondary endpoint. In OCEANIC-STROKE, ischemic stroke within 90 days occurred in 3.0% of patients on asundexian and 3.5% on placebo, a hazard ratio of 0.84 with a 95% confidence interval of 0.69 to 1.02 and a P value of 0.08. Because that endpoint missed, formal hierarchical testing of the lower-ranked endpoints stopped there. Over the full follow-up, by contrast, the hazard ratio for ischemic stroke was 0.74, with an interval of 0.65 to 0.84, well clear of no effect.

FROM THE DOCUMENT
was increased with Sunharbo in patients older than 75 years and during the first 90 days of treatment
The FDA's own wording on major bleeding with Sunharbo. Source: FDA press announcement: FDA Approves First Factor XIa Inhibitor to Help Prevent Stroke in Certain Adults at High Risk (Nov. 9, 2026)
ISCHEMIC STROKE, ASUNDEXIAN VS PLACEBO (HR)
Figure data
ResultValue (HR)Lower boundUpper bound
Full follow-up0.740.650.84
First 90 days0.840.691.02
Over the whole trial the interval clears no effect. In the first 90 days it does not. Source: Practical Neurology: OCEANIC-STROKE results published in NEJM (Sharma et al., 2026)

The long game met its target. The first 90 days did not. During a patient’s first three months on the drug, the trial cannot show with statistical confidence that it is preventing strokes, and in that same stretch the agency reports more major bleeding. That does not prove those weeks are a net loss, since a P value of 0.08 is an inconclusive result rather than a demonstrated failure. It does mean the approval rests on a benefit that builds over time, while the bleeding signal the FDA flagged sits at the start, among the patients most recently discharged from the stroke unit.

What 26 percent buys

The FDA approved Sunharbo as a 50 mg once-daily tablet, taken with antiplatelet drugs, for adults who have just had a mild-to-moderate non-cardioembolic ischemic stroke (a clot that did not start in the heart) or a high-risk transient ischemic attack. It is the first factor XIa inhibitor to be approved, and it arrived with Fast Track and Priority Review designations. The headline figure in the agency’s announcement is a 26% reduction in ischemic stroke against placebo over a median treatment of 14.2 months.

That figure is accurate, and it is a relative one. OCEANIC-STROKE randomized 12,327 adults at 702 centers in 37 countries, and ischemic stroke occurred in 6.2% of the asundexian group against 8.4% on placebo. That is an absolute difference of 2.2 percentage points over a median follow-up of 567 days. Averaged across the trial, roughly 45 patients would have to be treated over that span to prevent one ischemic stroke. In a disease that takes away speech and movement, that benefit is worth having. It also sounds smaller than 26%, which is why the press release was written in percentages. The composite of cardiovascular death, heart attack or stroke fell too, from 11.1% to 9.2%.

Bleeding is where the factor XI class was supposed to separate itself from older anticoagulants. ISTH major bleeding came in at 1.9% on asundexian and 1.7% on placebo, a hazard ratio of 1.10 that did not reach significance, and the FDA’s word for it was “similar.” The confidence interval runs from 0.85 to 1.44, so the trial remains compatible with a major-bleeding risk as much as 44% higher. Major or clinically relevant nonmajor bleeding was 5.5% against 5.0%, and the agency says gastrointestinal bleeding was more common on the drug. What the trial showed was no significant overall increase, which is not the same as showing no increase, and the over-75 finding on the label is the difference written down.

A molecule with a record

The theory behind the drug is older than any of its trials. Factor XI helps a clot grow, and the hope has long been that blocking it would stop dangerous clots while sparing the body’s ability to seal a wound. HCPLive describes asundexian as “designed to reduce pathologic thrombus formation while largely preserving hemostasis.”

The record has been less obliging. In OCEANIC-AF, asundexian went head to head with apixaban, the anticoagulant better known as Eliquis, in 14,810 people with atrial fibrillation, and in 2023 the independent data monitoring committee recommended stopping it early. When the results were presented in 2024, 98 patients on asundexian had suffered a stroke or systemic embolism against 26 on apixaban. Principal investigator Manesh Patel’s verdict was that “asundexian was found to be inferior.” It did bleed less, with 17 major bleeds against 53, so in that trial the factor XI bargain worked out to less bleeding and less protection. Sunharbo’s label says it is not approved for stroke prevention in atrial fibrillation.

OCEANIC-AF: STROKE OR SYSTEMIC EMBOLISM (patients)
Figure data
GroupValue (patients)
Asundexian98
Apixaban26
The atrial fibrillation trial that was stopped early. Source: American College of Cardiology: OCEANIC-AF, asundexian vs. apixaban in atrial fibrillation (ESC 2024)

The stroke program had stumbled once before that. In the phase 2b PACIFIC-Stroke trial, 1,808 patients took one of three doses or placebo, and asundexian “did not reduce the composite of covert brain infarction or ischemic stroke and no dose response could be shown.” What survived was an exploratory signal: the 50 mg dose reduced recurrent symptomatic ischemic strokes and TIAs, particularly in patients with atherosclerosis. Ashkan Shoamanesh, a scientist at the Population Health Research Institute, said the promising results “require validation in an adequately-powered phase 3 randomised trial.” OCEANIC-STROKE was that trial, and its entry rules leaned toward the patients the signal had favored. Each participant needed a nonlacunar infarct on imaging, a history of atherosclerosis, or atherosclerotic plaque on cerebrovascular imaging. That is sensible design. It also means the population that won in phase 3 resembles the subgroup that looked best in a trial that missed its primary endpoint.


The label carries the usual fine print for a new anticoagulant. Sunharbo is not recommended with drugs that are combined P-gp and strong CYP3A4 inducers, carbamazepine among them, or in moderate or severe liver impairment. It has to be withheld at least 24 hours before elective surgery and for at least 24 hours after thrombolysis. The FDA’s announcement names no reversal agent, though the same announcement flags the drug’s early bleeding.

Bayer, which developed the drug, was less restrained than the agency. Sara Hegab, its U.S. medical affairs vice president for stroke and thrombosis, called the NEJM publication “a class-defining contribution to the field of secondary stroke prevention.” Commercially, class-defining means first to market.

The editorial that ran with the trial in NEJM was warmer than skeptical but more careful than Bayer. Marion Boulanger called asundexian “a very promising option for secondary prevention after noncardioembolic ischemic stroke or TIA,” and in the same piece wrote that “further studies of factor XI inhibitors with longer follow-up (e.g., 5 or 10 years) are necessary” to see whether the net clinical benefit holds. The FDA did not wait for that answer. Patients will start Sunharbo where its trial evidence is thinnest, in the first 90 days, the period the agency itself singled out for more major bleeding.

Sources

  1. FDA press announcement: FDA Approves First Factor XIa Inhibitor to Help Prevent Stroke in Certain Adults at High Risk (Nov. 9, 2026)
  2. Practical Neurology: OCEANIC-STROKE results published in NEJM (Sharma et al., 2026)
  3. NeurologyLive: Phase 3 OCEANIC-STROKE data, design and investigator and Bayer statements
  4. American College of Cardiology journal scan: OCEANIC-STROKE and the accompanying NEJM editorial
  5. HCPLive: Asundexian receives FDA approval, label restrictions and subgroup bleeding
  6. American College of Cardiology: OCEANIC-AF, asundexian vs. apixaban in atrial fibrillation (ESC 2024)
  7. Population Health Research Institute: PACIFIC-Stroke phase 2b results (ESC 2022)