Bristol Myers Squibb and Johnson & Johnson have received definitive Phase 3 confirmation that milvexian failed to reduce major cardiovascular events when added to standard antiplatelet therapy following an acute coronary syndrome, delivering a major setback to one of the pharmaceutical industry’s most closely watched attempts to create a safer generation of anticoagulants.
The LIBREXIA ACS trial randomized 14,194 patients across 893 sites in 44 countries to milvexian 25 mg twice daily or placebo following a recent heart attack or related acute coronary syndrome event. Cardiovascular death, myocardial infarction or ischemic stroke occurred in 5.4% of patients receiving milvexian compared with 5.1% receiving placebo, producing a hazard ratio of 1.05 and no evidence of cardiovascular benefit.
Why did Bristol Myers Squibb and Johnson & Johnson test milvexian after a heart attack?
Patients remain at elevated risk of another clot-driven cardiovascular event after acute coronary syndrome even when physicians use modern stents, statins and dual antiplatelet therapy.
Adding stronger anticoagulation appears logically attractive, but conventional anticoagulants can increase serious bleeding when layered on top of antiplatelet medicines. That has created a long-running search for a treatment capable of suppressing dangerous thrombosis without substantially impairing the body’s ability to stop normal bleeding.
Factor XIa emerged as one of the most promising targets.
People with naturally lower factor XI activity appear to have lower thrombotic risk without suffering the same severe spontaneous bleeding associated with deficiencies of some other coagulation proteins. Drug developers therefore hypothesized that inhibiting factor XIa might uncouple thrombosis prevention from major bleeding.
Milvexian is one of the leading oral medicines built around that idea.
What went wrong in the LIBREXIA ACS Phase 3 trial?
The trial did not fail because of excessive fatal bleeding. It failed because the medicine did not prevent enough cardiovascular events.
After a median follow-up around one year, 384 patients receiving milvexian experienced the primary endpoint compared with 365 receiving placebo. The difference was not statistically significant and numerically favored placebo rather than milvexian.
The study had actually been stopped for futility in November 2025 after an independent monitoring committee concluded during a prespecified interim analysis that the trial was unlikely to achieve its primary efficacy objective. Bristol Myers Squibb and Johnson & Johnson disclosed that decision at the time but continued following the remaining milvexian program.
The complete results presented at European Society of Cardiology Congress 2026 and published in the New England Journal of Medicine now quantify exactly why the program was stopped.
Did milvexian at least avoid serious bleeding?
Yes, and that is why the overall factor XIa story is not finished.
Intracranial or fatal bleeding occurred in approximately 0.3% of patients in both the milvexian and placebo groups. The principal safety comparison therefore showed no significant increase in the most feared bleeding outcomes.
Laboratory testing also showed the expected anticoagulant effect, indicating that milvexian was biologically active.
The problem was that biological anticoagulation did not translate into fewer cardiovascular deaths, myocardial infarctions or ischemic strokes in this particular setting.
That result creates a difficult but scientifically interesting conclusion: factor XIa inhibition may indeed achieve unusually favorable bleeding characteristics, but that advantage is commercially useful only if the drug also prevents clinically important thrombotic events.
Does the failure mean milvexian development is over?
No.
Two large Phase 3 programs remain critical to the drug’s future. LIBREXIA AF is evaluating milvexian for stroke prevention in atrial fibrillation, while LIBREXIA STROKE is studying the medicine in secondary stroke prevention.
Those populations are biologically and therapeutically different from patients immediately following acute coronary syndrome.
In the ACS trial, patients were already receiving antiplatelet therapy after a recent coronary event. It is possible that adding factor XIa inhibition offered too little incremental protection against platelet-rich coronary thrombosis.
Atrial fibrillation, by comparison, involves blood stasis and thrombus formation within the heart, an environment in which anticoagulation plays a much more established therapeutic role.
Researchers presenting LIBREXIA ACS emphasized that the negative result cannot automatically be transferred to the remaining trials because the diseases, background therapies and endpoints differ.
Why does the factor XI drug class still matter after this failure?
Bleeding remains one of the principal reasons patients who need anticoagulation receive insufficient treatment or no treatment at all.
Current oral anticoagulants targeting factor Xa or thrombin have transformed stroke prevention, particularly in atrial fibrillation, but physicians still balance every reduction in clot risk against an increase in bleeding risk.
A successful factor XIa inhibitor could therefore be commercially enormous if it delivers comparable protection with meaningfully less bleeding.
Bristol Myers Squibb had previously described milvexian as a potential multibillion-dollar opportunity across acute coronary syndrome, stroke and atrial fibrillation. The ACS indication alone represented an estimated population of roughly two million patients across major U.S. and European markets in earlier company materials.
The failure removes one major branch from that commercial thesis.
Could the failure actually teach researchers where factor XI inhibition works best?
Possibly.
Drug-development failures sometimes invalidate an entire mechanism. Others clarify where that mechanism should and should not be used.
LIBREXIA ACS suggests that suppressing factor XIa on top of standard antiplatelet therapy after an acute coronary event does not add enough efficacy to justify treatment at the tested dose.
If the atrial-fibrillation or stroke trials succeed, the interpretation could become more specific: factor XI inhibition may work best in thrombotic settings where coagulation pathways play a larger role than platelet activation.
If those trials also fail, however, the industry will face a much harder question about whether the attractive human genetics and bleeding biology of factor XI can be converted into practical drug efficacy.
What should the pharmaceutical industry watch next?
LIBREXIA AF and LIBREXIA STROKE now carry substantially more strategic weight.
Positive results from either program could preserve milvexian as a major anticoagulation franchise and demonstrate that the ACS failure was disease-specific. Negative results would challenge one of the strongest mechanistic hypotheses in cardiovascular drug development.
LIBREXIA ACS therefore matters even though companies had already disclosed the futility decision months ago. The final 14,194-patient dataset shows that milvexian achieved the reassuring half of its promise, avoiding an obvious increase in intracranial or fatal bleeding, while failing the part that ultimately determines whether a cardiovascular drug can succeed: preventing more cardiovascular events than existing therapy alone.
