Daiichi Sankyo’s Edoxaban Falls Short in ENRICH-AF as Major Bleeding More Than Doubles

Daiichi Sankyo’s edoxaban failed to produce a statistically significant reduction in stroke or systemic embolism in ENRICH-AF, a randomized trial involving patients with atrial fibrillation who had survived an intracranial hemorrhage. The result was accompanied by a clear safety penalty: major bleeding occurred in 11.6% of patients assigned to edoxaban, compared with 5.2% in the no-anticoagulation group.

The finding matters because ENRICH-AF tested anticoagulation in a population where the usual stroke-prevention trade-off is unusually difficult. Atrial fibrillation raises the risk of ischemic stroke, but a previous brain bleed makes renewed anticoagulation more hazardous. In this trial, reductions in some clot-related events were not enough to offset the increase in hemorrhagic outcomes.

ENRICH-AF Missed Its Primary Efficacy Endpoint

The study enrolled 948 patients with high-risk atrial fibrillation and a previous intracranial hemorrhage. It was conducted at 174 sites in 20 countries, with participants randomly assigned in equal proportions to edoxaban or no anticoagulation. Edoxaban was given at 60 mg once daily, with a reduction to 30 mg when required by the drug’s label. The comparison group could receive no antithrombotic therapy or a single antiplatelet drug, based on the treating clinician’s decision.

After an average follow-up of 28 months, the primary endpoint of stroke, including ischemic and hemorrhagic stroke, or systemic embolism occurred in 11.8% of the edoxaban group and 12.8% of the comparison group. The hazard ratio was 0.88, with a 95% confidence interval of 0.61 to 1.26 and a p value of 0.48. The European Society of Cardiology reported the results after their presentation in a Hot Line session at ESC Congress 2026 on August 29.

A hazard ratio below 1 pointed toward fewer primary events with edoxaban, but the confidence interval crossed 1 and the difference was not statistically significant. The component results also moved in different directions. The ESC said ischemic stroke and myocardial infarction were significantly reduced with edoxaban, but that advantage was offset by an almost three-fold excess in hemorrhagic stroke. The overall result therefore did not establish a net efficacy benefit for routine edoxaban use across this high-risk post-hemorrhage population.

Major Bleeding More Than Doubled

Major bleeding, defined by the International Society on Thrombosis and Haemostasis, was the primary safety outcome. It occurred in 11.6% of patients receiving edoxaban and 5.2% of those assigned to no anticoagulation. The hazard ratio was 2.23, with a 95% confidence interval of 1.39 to 3.59 and a p value below 0.001. The absolute difference was 6.4 percentage points, and the hazard ratio showed more than a doubling of major bleeding risk.

Safety concerns had already affected the trial before the final readout. Following a Data and Safety Monitoring Board review in 2023, investigators stopped enrolling additional patients with lobar intraparenchymal hemorrhage or convexity subarachnoid hemorrhage because of safety concerns. That change is important to the interpretation of ENRICH-AF because the trial was never simply a test of anticoagulation in a typical atrial fibrillation population. It focused on people with a history that placed them at unusually high risk if bleeding recurred.

ENRICH-AF used an investigator-initiated, open-label design with blinded endpoint assessment and was event driven. Participants had a mean age of 77 years, and 39% were women. The Population Health Research Institute was the study sponsor, while the ESC said the trial was supported by grants-in-aid from Daiichi Sankyo and the institute. Those features do not change the randomized comparison, but they are relevant context when evaluating the evidence and its intended clinical scope.

Principal investigator Ashkan Shoamanesh said the findings do not support edoxaban for unselected patients with atrial fibrillation after intracranial hemorrhage and instead point toward individualized decision-making. That distinction is central. The result does not establish that every patient with a previous brain bleed should avoid anticoagulation, but it does show that a broad strategy of routinely giving edoxaban to this population did not achieve the trial’s primary efficacy objective and produced substantially more major bleeding.

Edoxaban Remains a Major Daiichi Sankyo Product

For Daiichi Sankyo, ENRICH-AF is a negative clinical readout for a drug that remains commercially important. Edoxaban is sold as Lixiana in markets including Europe and Japan and as Savaysa in the United States. Existing approvals cover stroke and systemic embolism prevention in adults with nonvalvular atrial fibrillation who have relevant risk factors, among other uses. ENRICH-AF addressed a much narrower question involving patients who had already experienced intracranial hemorrhage, so the result is not a readout on the broader approved atrial fibrillation population.

Daiichi Sankyo’s latest quarterly reference data, released July 31, showed global edoxaban revenue of 90.7 billion yen in the first quarter of fiscal 2026, down 0.4% from a year earlier. Lixiana revenue in Europe rose 7.7% to 49.2 billion yen, while Japanese Lixiana revenue fell 12.2% to 33.0 billion yen. The figures underline the drug’s continuing place in Daiichi Sankyo’s product base.

ENRICH-AF does not by itself change those existing approvals. The more direct implication is evidentiary: the trial failed to support routine edoxaban use in an especially vulnerable group where clinicians have had limited randomized data. For investors, that makes the readout more relevant to the drug’s potential use in this difficult post-hemorrhage setting than to its established role across the broader atrial fibrillation market.

Further evidence is still developing. The ESC highlighted the ongoing ASPIRE trial, which is comparing apixaban with aspirin, as well as the planned COCROACH individual-participant-data meta-analysis of randomized trials. Those results may help determine whether specific clinical characteristics can identify patients whose ischemic-stroke benefit from anticoagulation outweighs the danger of another intracranial bleed. ENRICH-AF makes that selection problem harder to ignore: in an unselected high-risk population, the expected protection against clot-related events came with too much bleeding to produce a statistically significant overall benefit.