On October 26, 2026, China's National Medical Products Administration (NMPA) approved Bayer's Class 1 innovative drug asundexian (Chinese trade name: 阿舒克森片) through the priority review and approval pathway, for reducing the risk of stroke in adult patients with ischemic stroke or high-risk TIA, excluding those with cardioembolic stroke. It is the first activated coagulation factor XI (FXIa) inhibitor approved anywhere in the world, and China is the first country to approve it.
For the development side, the value of this news goes beyond a single product reaching the market. FXIa is a pathway that has been under discussion for more than a decade and long remained at the level of theory. It has now produced completely contradictory results in two Phase III studies of ten thousand patients each. From the standpoint of trial design and execution, this article analyzes why the same molecule won in one population and lost in another, and what this design can teach ongoing antithrombotic programs.
1. The Event Itself: The First Positive Phase III Support for a Target
The pivotal study supporting approval was OCEANIC-STROKE, whose results were presented at the International Stroke Conference in February 2026 and published in full in the New England Journal of Medicine on April 16. Core parameters are as follows:
| Parameter | Value |
|---|---|
| Design | Phase III, randomized, double-blind, placebo-controlled, event-driven, 1:1 allocation |
| Sample Size | 12,327 patients (asundexian 6,162 / placebo 6,165) |
| Centers & Countries | 702 sites across 37 countries and regions |
| Enrollment & Follow-up | January 2023 to February 2025; median follow-up 567 days (about 19 months) |
| Dosing | Asundexian 50 mg orally once daily, added on top of single- or dual-antiplatelet therapy |
| Primary Efficacy Endpoint | First ischemic stroke: 6.2% vs 8.4%, csHR 0.74 (95% CI 0.65–0.84), P<0.001; by 1-year Aalen-Johansen estimate 5.1% vs 7.0%, absolute reduction 1.9%, NNT approximately 53 |
| Primary Safety Endpoint | ISTH major bleeding: 1.9% vs 1.7%, HR 1.10 (95% CI 0.85–1.44), P=0.46 |
| Key Secondary Endpoints | CV death / MI / stroke 9.2% vs 11.1%, HR 0.83; any stroke 6.6% vs 8.8%, HR 0.74; symptomatic intracranial hemorrhage 0.7% vs 0.6% |
This is the first Phase III study to complete successfully in the FXIa inhibitor field. The core hypothesis for this pathway had been the "decoupling of thrombosis and hemostasis" — that inhibiting FXIa could weaken the amplification of pathological thrombi while leaving normal hemostasis largely unaffected. OCEANIC-STROKE is the first time this hypothesis was fully validated in a large, event-driven study.
2. One Molecule, Two Opposite Conclusions: Population Choice Decides Outcome Earlier Than Target Validation
To understand the boundaries of this approval, one has to return to OCEANIC-AF, presented at the European Society of Cardiology congress in August 2024. Same molecule, same dose, but the population switched to high-risk atrial fibrillation patients and the comparator switched to apixaban. The study was terminated early by the independent data monitoring committee at a median follow-up of about 155 days (stop recommended on November 19, 2023).
| Dimension | OCEANIC-STROKE | OCEANIC-AF |
|---|---|---|
| Population | Non-cardioembolic ischemic stroke or high-risk TIA | Atrial fibrillation with high stroke risk |
| Sample Size | 12,327 | 14,810 |
| Therapeutic Setting | Added on top of antiplatelet therapy; placebo comparator | Substitute for anticoagulation; apixaban comparator |
| Thromboembolic Events | 6.2% vs 8.4%, HR 0.74 | 1.3% vs 0.4%, HR 3.79 (2.46–5.83) |
| Major Bleeding | 1.9% vs 1.7%, HR 1.10 | 0.2% vs 0.7%, csHR 0.32 (0.18–0.55) |
| Outcome | Met both primary efficacy and safety endpoints | Terminated early for lack of efficacy |
In OCEANIC-AF, major bleeding with asundexian was significantly lower than with apixaban (0.2% vs 0.7%), while stroke and systemic embolism occurred roughly three times as often as in the control arm. Some published analyses hold that a 50 mg once-daily dose may not sufficiently suppress FXIa activity in the AF setting. Other investigators point out that even at trough, this dose still maintains roughly 92% inhibition of FXIa, suggesting the problem may not lie in exposure alone — the thrombus formation mechanism in AF patients may itself differ from that in atherosclerotic populations. In OCEANIC-STROKE, the same dose showed a positive net benefit.
The development implication from this contrast is direct: the FXIa pathway itself is not invalidated by OCEANIC-AF. What failed was the specific proposition of "replacing anticoagulation with FXIa inhibition." AF-related stroke arises from fibrin-rich thrombi formed by stasis within the atrium, and protecting against it requires anticoagulation of sufficient intensity. The residual risk of non-cardioembolic stroke, by contrast, comes from platelet-dependent thrombosis on a background of atherosclerosis — adding a mechanistically complementary pathway on top of antiplatelet therapy is what constitutes a reasonable combination hypothesis.
In other words, target validation came later than scenario validation. Before entering a pivotal Phase III, a program needs to answer not "does this target act on thrombi" but "in this patient population's thrombus formation mechanism, does this pathway hold an irreplaceable position, and is its relationship to the drugs already in use complementary or substitutive."
3. Four Reusable Design Lessons
Design Principles
- Additive control rather than active control. The OCEANIC-STROKE comparator was placebo, but both arms received physician-directed single- or dual-antiplatelet therapy, with about 63% of patients planned for DAPT. This design lets the study answer directly whether there is incremental benefit on top of existing standard of care, without having to litigate non-inferiority margins as an active-comparator design requires. For antithrombotic drugs with a complementary mechanism, this approach is both more economical and clearer in evidentiary terms.
- Deliberately relaxed enrollment in exchange for external validity. The NIHSS upper limit was set at 15, markedly broader than previous stroke secondary-prevention studies, which typically enrolled only minor strokes. Patients who had received IV thrombolysis or mechanical thrombectomy were eligible 24 hours after the end of treatment, accounting for roughly 27% of the cohort. High-risk TIA was defined by ABCD2 scores of 6 to 7. The cost of relaxed enrollment is pressure on event accrual and sample size; the benefit is a result more representative of real hospitalized populations. This trade-off needs to be aligned across statistics, clinical and operations teams at the protocol stage, not adjusted retroactively once enrollment slows.
- Standardized bleeding definitions for the primary safety endpoint. The primary safety endpoint used the ISTH major bleeding definition, with symptomatic intracranial hemorrhage (0.7% vs 0.6%), bleeding at other sites and other subcategories listed separately among the secondary endpoints. For antithrombotic drugs, differences in bleeding endpoint definitions can be enough to reverse the direction of the conclusion. Adopting internationally standardized definitions is both a regulatory requirement and a precondition for cross-study comparison. Endpoint adjudication requires independence, blinding and uniform operating procedures — underestimating this workload is a common risk in antithrombotic programs.
- Event-driven design with sufficiently long follow-up. The study used event-driven enrollment with a median follow-up of 567 days (interquartile range 377 to 729 days). Stroke recurrence is a low-incidence endpoint, and the 1-year absolute reduction was only 1.9 percentage points, requiring enough events for stable detection; meanwhile, safety signals for antithrombotic drugs often only emerge after prolonged exposure. The assumptions about event rate made at protocol design, and the tolerance for follow-up duration, directly determine whether the study can close within a reasonable timeframe.
4. Three Unavoidable Execution Questions
Enrollment Window and Screening Throughput
Entry required randomization within 72 hours of symptom onset, and dosing to start within 4 hours of randomization. Screening, informed consent, imaging assessment, randomization and first dose therefore have to form an hour-level closed loop inside the hospital, involving coordinated multi-point workflows across emergency, imaging, neurology and pharmacy. For a study spanning 702 sites across 37 countries, differences in site workflow are the primary source of variation in enrollment speed and protocol compliance.
Consistency of Etiologic Subtype Interpretation
The enrolled population covered large-artery atherosclerosis (about 43%), small-vessel occlusion (about 22%) and undetermined cause (about 30%), with a clear exclusion of patients with a history of atrial fibrillation or other cardioembolic sources requiring anticoagulation. Such subtyping depends on combined interpretation of imaging and clinical information. Inter-center consistency directly affects population purity, which in turn affects how interpretable the endpoint results are. Operational definitions of the subtyping criteria, center training and imaging review mechanisms all need to be fixed at study start-up.
Alignment with Existing Stroke Center Workflows
The rule permitting enrollment 24 hours after thrombolysis or thrombectomy is directly coupled to the acute-phase treatment pathway. Most enrolled patients were already on antiplatelet medication, making documentation, verification and adherence management of concomitant medication a foundational data-quality item. The data-management burden of such studies concentrates along three lines: imaging, concomitant medication and bleeding events.
5. Three Implications for Ongoing Programs
- The population hypothesis should be tested before the dose hypothesis. OCEANIC-AF and OCEANIC-STROKE used the same dose and reached opposite conclusions, showing that in the wrong therapeutic scenario, adjusting the dose may not rescue the program — while in the right scenario, the same dose can hold up. The critical decision point in early programs is to first determine what background therapy the product is layered onto and what thrombus mechanism it targets, and only then discuss exposure.
- Safety endpoints must be defined at protocol stage. "No increase in major bleeding" is one of the core supports for this approval, and the credibility of that conclusion depends entirely on standardized endpoint definitions and independent adjudication. Treating bleeding endpoints as a post-hoc analysis carries excessive risk.
- The registration window for global synchronized development is shortening. This product's marketing application was accepted by the NMPA on April 30, 2026 and granted priority review, with approval on October 26 — a span of roughly six months, making China the first country to approve. New drug applications in the United States and Japan were both accepted in May 2026 (the FDA also granted priority review) and remain unapproved. For products with clearly unmet needs, the actual efficiency of the Chinese regulatory channel deserves to be factored into early registration strategy, rather than waiting until the first global filing is complete before following up.
Validation of the FXIa pathway took more than a decade, and what finally made it work was not a stronger molecule but a more accurate definition of the scenario. For any target whose "mechanism sounds plausible," this case is worth revisiting at project initiation and at key decision milestones.
Sources: NMPA announcement, October 26, 2026; OCEANIC-STROKE results (presented at the International Stroke Conference in February 2026, published in full in the New England Journal of Medicine on April 16, 2026); OCEANIC-AF results (presented at the European Society of Cardiology congress in August 2024, published in the New England Journal of Medicine in the same period); Bayer public information. This article is for information sharing on research design and execution only, and does not constitute dosing advice or medical opinion.