Blood Thinners and Brain Bleeds: New Guideline Rewrites Emergency Reversal Rules

When a patient on blood thinners suffers a brain hemorrhage, every minute counts, and the choice of reversal drug can mean the difference between recovery and catastrophe. A decade after issuing its first comprehensive guidance, the Neurocritical Care Society together with the Society of Critical Care Medicine has published a focused update that fundamentally reshapes how clinicians should treat life-threatening intracranial hemorrhage in adults taking antithrombotic medications. The new guideline, published open access in the journal Neurocritical Care, delivers eight conditional recommendations built on a rigorous systematic review and meta-analysis of the evidence that has accumulated since 2016, and its headline finding is striking: for patients bleeding on oral factor Xa inhibitors, the widely marketed antidote andexanet alfa should be passed over in favor of the older, cheaper, and apparently safer four-factor prothrombin complex concentrate.

The stakes of this question are enormous. Antithrombotic drugs, including anticoagulants and antiplatelet agents, are prescribed to millions of people for atrial fibrillation, venous thromboembolism, and cardiovascular disease, and they dramatically increase the risk and severity of intracranial hemorrhage. Compared with patients who bleed while not taking these medications, those on antithrombotics face higher rates of hematoma expansion, worse functional outcomes, and greater mortality. Hematoma expansion is the central villain in this story: every additional milliliter of blood within the brain parenchyma carries roughly a five percent increased risk of death or dependency, and approximately a third of patients who present within six hours of symptom onset will experience significant growth of their hemorrhage. Rapid reversal of anticoagulation is therefore one of the most consequential decisions in emergency neurology.

The guideline panel, an interdisciplinary task force of ten specialists spanning neurocritical care, trauma surgery, hematology, nursing, and pharmacy, framed five clinical questions using the Population, Intervention, Comparator, Outcome framework. Three questions updated prior recommendations while two were entirely new. The panelists independently rated outcomes by importance, elevating mortality, functional outcome, and thrombosis to critical status, and applied the Grading of Recommendations, Assessment, Development, and Evaluation methodology to synthesize evidence from randomized trials and observational cohorts. Notably, the panel included a public member recruited through the Anticoagulation Forum to represent the patient and family perspective, and all members were vetted for conflicts of interest before voting.

The most consequential recommendation concerns the showdown between andexanet alfa and prothrombin complex concentrate. Andexanet alfa is an inactive recombinant decoy form of human factor Xa that sponges up factor Xa inhibitors and restores thrombin generation. It was not commercially available when the 2016 guideline was written, and its subsequent rise was propelled largely by the ANNEXA-4 trial, a single-arm study of 352 patients in which 82 percent achieved excellent or good hemostasis. But the new analysis, drawing on 21 studies encompassing 4,938 patients, tells a more sobering story. The sole randomized trial, ANNEXA-I, compared andexanet alfa against usual care in factor Xa inhibitor-associated intracerebral hemorrhage and found no significant mortality benefit: 27.8 percent of andexanet recipients died versus 25.5 percent of those receiving usual care. Andexanet did improve hemostatic efficacy, with 67 percent achieving the composite hemostasis endpoint compared with 53.1 percent under usual care, and it produced less very large hematoma expansion. Yet hemostasis is a surrogate outcome, and the panel prioritized patient-centered endpoints instead.

What tipped the scales decisively was thrombosis. Meta-analysis revealed a significant increase in thrombotic events with andexanet alfa, with a relative risk of 1.37 overall, and in the spontaneous hemorrhage subgroup the risk nearly doubled at a relative risk of 1.99. These events were predominantly arterial, most commonly acute ischemic strokes identified on neuroimaging, and even small asymptomatic infarcts may erode long-term functional recovery. Andexanet alfa appears prothrombotic through two distinct mechanisms: reversal of anticoagulation itself, and direct inhibition of tissue factor pathway inhibitor. These findings contributed to the United States Food and Drug Administration declining AstraZeneca’s application for full approval, and the manufacturer voluntarily withdrew the drug from the US market as of December 2025, though it remains available internationally. The panel therefore issued a conditional recommendation favoring four-factor prothrombin complex concentrate over andexanet alfa for both spontaneous and traumatic factor Xa inhibitor-associated hemorrhage, the latter based on very low certainty evidence.

The guideline also delivers a nuanced verdict on platelet transfusion, long a reflexive response to bleeding on antiplatelet drugs like aspirin and clopidogrel. For patients with spontaneous intraparenchymal hemorrhage who do not require surgery, the panel recommends against platelet transfusion, a position anchored by the PATCH trial, a high-quality randomized study showing that transfusion offered no benefit in mortality, functional outcome, or thrombosis, and in fact was associated with worse functional outcomes. The picture flips for surgical patients: a single randomized trial from China found that aspirin users undergoing craniotomy for basal ganglia hemorrhage who received perioperative platelet transfusion had lower mortality, reduced postoperative hemorrhage, and better six-month functional outcomes. The panel accordingly suggests platelet transfusion for aspirin-treated patients requiring neurosurgery, while explicitly limiting this recommendation to aspirin because no data exist for P2Y12 inhibitors in this setting. For traumatic intracranial hemorrhage, where ten retrospective studies produced contradictory and confounded results, the panel declined to make any recommendation at all.

Desmopressin, a vasopressin analog that releases factor VIII and von Willebrand factor and may enhance platelet procoagulant activity, fared no better. The 2016 guideline had suggested considering a single dose for antiplatelet-associated hemorrhage, but the new analysis, including the DASH phase 2 feasibility trial and nine observational studies, found no effect on hematoma expansion, mortality, or functional outcome, with all confidence intervals crossing unity. The panel made no recommendation, while emphasizing the urgent need for adequately powered randomized trials of a treatment that remains biologically plausible for a condition with no proven therapy. Similarly, the panel could not issue guidance on whether to reverse anticoagulation in patients with very small intraparenchymal hemorrhages, a genuinely difficult clinical dilemma, since no studies directly address it. The decision hinges on factors such as time from onset to imaging, hemorrhage location, the specific anticoagulant and timing of the last dose, and imaging markers like the CT spot sign that predict expansion.

One genuinely new recommendation embraces technology over therapeutics. Viscoelastic hemostatic assays, including thromboelastography and rotational thromboelastometry, are point-of-care whole blood tests that render real-time coagulation dynamics rather than static snapshots. The panel suggests using these assays to guide treatment of coagulopathy in traumatic intracranial hemorrhage when available, a conditional recommendation resting on very low certainty evidence but supported by a mortality signal in the severe traumatic brain injury subgroup of one randomized trial, with a relative risk of 0.59. Because the tests carry minimal inherent risk compared with medications or transfusions, the panel judged the balance favorable even with sparse data, though it acknowledged that treatment protocols vary widely between centers and that standardized algorithms remain lacking.

The panel was candid about the limitations pervading this literature. Observational studies comparing andexanet with prothrombin complex concentrate are vulnerable to selection bias, since patients assigned to usual care were often treated later, had higher blood pressures, or were perceived as having better prognoses, patterns that could artificially favor andexanet. Formulations of prothrombin complex concentrate differ in factor content, some contain heparin, and none carries a labeled indication for factor Xa inhibitor reversal, while optimal dosing, historically 50 units per kilogram, was never established on robust data. The panel also advised against combining andexanet alfa with prothrombin complex concentrate unless benefit clearly outweighs the prothrombotic risk, and flagged a practical pitfall: patients transferred between hospitals may arrive without documentation of reversal agents already given.

Perhaps the most telling conclusion is what the guideline could not say. Not a single strong recommendation emerged from the entire exercise, a reflection of how thin the evidence remains for one of the most feared complications of modern cardiovascular medicine. The panel laid out a research agenda calling for randomized trials of platelet transfusion in traumatic hemorrhage, adequately powered studies of desmopressin, direct comparisons of reversal versus observation in small hemorrhages, and standardized viscoelastic assay protocols. Until those trials arrive, clinicians now have a clearer, more skeptical map of the territory: favor prothrombin complex concentrate over andexanet, withhold platelets from non-surgical patients, give them to aspirin users heading for the operating room, and let real-time coagulation testing steer trauma resuscitation. In a field where overtreatment has proven as dangerous as undertreatment, knowing what not to do may be the guideline’s most valuable contribution.

Subject of Research: Guideline-based reversal of antithrombotic medications in adults with acute intracranial hemorrhage

Article Title: Treatment of Antithrombotic-Associated Intracranial Hemorrhage in Adults: A Focused Guideline Update from the Neurocritical Care Society and the Society of Critical Care Medicine

Article References: Vallejo, M. C., Kennedy, L. S., Dengler, B., Barthol, C., Belley-Cote, E. P., Burns, J. D., Chau, V. K., Cordonnier, C., Cuker, A., Honarmand, K., Mahmoud, S. H., Wilcox, M. E., & Kumar, M. A. (2026). Treatment of Antithrombotic-Associated Intracranial Hemorrhage in Adults: A Focused Guideline Update from the Neurocritical Care Society and the Society of Critical Care Medicine. Neurocritical Care. https://doi.org/10.1007/s12028-026-02601-4

Image Credits: AI Generated

DOI: 10.1007/s12028-026-02601-4

Keywords: intracranial hemorrhage, antithrombotic reversal, andexanet alfa, prothrombin complex concentrate, platelet transfusion, desmopressin, factor Xa inhibitors, antiplatelet agents, viscoelastic hemostatic assays, GRADE methodology, neurocritical care, clinical guideline