What do the ANNEXA‑4 and ANNEXA‑I trials reveal about the efficacy and safety of andexanet alfa for reversing apixaban or rivaroxaban in life‑threatening bleeding, and what are the recommended weight‑based dosing regimens?

Andexanet Alfa Trials: ANNEXA-4 and ANNEXA-I

Andexanet alfa rapidly reverses apixaban and rivaroxaban anticoagulation with 80% achieving excellent or good hemostasis at 12 hours, but carries a significant 10% thrombotic event risk within 30 days, making it the preferred FDA-approved reversal agent for life-threatening bleeding despite the absence of placebo-controlled mortality data. 1

ANNEXA-A and ANNEXA-R: Pharmacodynamic Foundation

These randomized, placebo-controlled trials in healthy volunteers established andexanet's mechanism of action:

• Apixaban reversal (ANNEXA-A): Andexanet reduced anti-FXa activity by 94% versus 21% with placebo and increased thrombin generation for up to 12 hours 1
• Rivaroxaban reversal (ANNEXA-R): Andexanet reduced anti-FXa activity by 92% versus 18% with placebo with similar thrombin generation restoration 1
• These phase 3 studies formed the basis for FDA approval in 2019, though they enrolled only healthy volunteers without actual bleeding 1

ANNEXA-4: The Pivotal Bleeding Trial

This prospective, open-label, single-arm study examined 479 patients (mean age 78 years, 46% female) with major bleeding within 18 hours of their last FXaI dose 1:

Patient Population and Bleeding Sites

• 51% were taking apixaban, 37% rivaroxaban, with 81% treated for atrial fibrillation 1
• 69% had intracranial hemorrhage, 23% gastrointestinal bleeding 1
• Patients with Glasgow Coma Scale <7 or hematoma volume >60 mL were excluded, limiting generalizability to the most severe ICH cases 1

Efficacy Outcomes

• Median anti-FXa activity decreased by 93% from baseline 1
• 80% of evaluable patients achieved excellent or good hemostasis at 12 hours using prespecified, bleeding-site-specific criteria 1
• Efficacy was analyzed only in 342 patients with baseline anti-FXa activity ≥75 ng/mL for apixaban/rivaroxaban 1

Critical Safety Concerns

• Thrombotic events occurred in 10% of participants within 30 days, with 16 events occurring despite parenteral thromboprophylaxis 1
• None of the patients who experienced thrombotic events had resumed oral anticoagulation, highlighting the importance of prompt anticoagulation resumption 1
• The absence of a control group fundamentally limits interpretation of both efficacy and safety findings—we cannot determine if outcomes differ from supportive care alone 1

ANNEXA-I: The First Randomized Controlled Trial

This randomized trial compared andexanet to usual care in 530 patients (mean age 79 years, 46% female) with intracranial hemorrhage presenting within 6 hours of symptom onset 1:

Study Design and Population

• Patients received their last FXaI dose (apixaban, rivaroxaban, or edoxaban) within 15 hours 1
• 85.5% of usual care patients received prothrombin complex concentrates (PCCs), making this effectively a comparison of andexanet versus PCC 1
• The trial was stopped early after interim analysis met efficacy criteria (P<0.031) 1

Primary Efficacy Endpoint (Effective Hemostasis at 12 Hours)

Defined as meeting ALL three criteria:

1. ≤35% hematoma volume expansion
2. NIH Stroke Scale score increase <7
3. No rescue therapy (surgery or PCC) between 3-12 hours post-randomization 1

67% of andexanet patients versus 53.1% of usual care patients achieved effective hemostasis (adjusted difference 13.4 percentage points; 95% CI 4.6-22.2; P=0.003) 1

Secondary Outcomes at 30 Days

• Mortality was similar: 27.8% with andexanet versus 25.5% with usual care (not statistically different) 1
• Functional outcomes were similar: 28.0% versus 30.9% achieved favorable functional status (modified Rankin Scale ≤3) 1
• Hematoma increase ≥12.5 mL or death within 12 hours: 11.1% with andexanet versus 16.8% with usual care 1

Thrombotic Risk Signal

Thromboembolic events were significantly more frequent with andexanet: 10.3% versus 5.6% (absolute increase 4.6 per 100 patients; 95% CI 0.1-9.2; P=0.048) 1

• Ischemic stroke specifically: 6.5% with andexanet versus 1.5% with usual care (absolute difference 5.0; 95% CI 1.5-8.8) 1
• This represents a clinically meaningful harm signal that must be weighed against the modest hemostatic benefit 1

Methodological Limitations

• Open-label design can bias use of cointerventions and outcome assessment 1
• Lack of standardized treatment within usual care group makes it difficult to determine the specific comparator 1
• The trial does not establish superiority for patient-centered outcomes (mortality, functional status) 1

Weight-Based Dosing Regimens

Andexanet is administered as an intravenous bolus followed by a 2-hour continuous infusion, with dosing dependent on the specific FXaI, dose, and timing of last intake 1, 2:

Low-Dose Regimen

• 400 mg IV bolus over 15 minutes (30 mg/min)
• Followed by 480 mg infusion over 2 hours (4 mg/min)
• Total dose: 880 mg 2

Indications:

• Apixaban ≤5 mg OR last dose ≥8 hours ago 2
• Rivaroxaban ≤10 mg OR last dose ≥8 hours ago 2

High-Dose Regimen

• 800 mg IV bolus over 30 minutes (30 mg/min)
• Followed by 960 mg infusion over 2 hours (8 mg/min)
• Total dose: 1,760 mg 2

Indications:

• Apixaban >5 mg OR last dose <8 hours ago 2
• Rivaroxaban >10 mg OR last dose <8 hours ago 2
• Edoxaban (any dose, off-label use) 2

Critical Dosing Considerations

• The reversal effect is transient—anti-FXa activity begins returning toward baseline approximately 2 hours after infusion completion 3
• Do not delay administration for laboratory confirmation in life-threatening bleeding 2, 3
• Anti-FXa activity >50 ng/mL is considered clinically significant for bleeding risk 2

Clinical Positioning Versus Alternatives

The American College of Cardiology suggests andexanet alfa is preferable to 4-factor PCC (4F-PCC) for major bleeding on oral FXaIs based on: 1

1. FDA approval based on protocolized prospective data (ANNEXA-4) with blind outcome adjudication 1
2. 4F-PCC data derive primarily from small observational studies with medical record review 1
3. Andexanet is the only FDA-approved treatment for apixaban/rivaroxaban-associated major bleeding 1

When Andexanet Is Unavailable

Administer 4F-PCC at 50 units/kg IV (maximum 4,000-5,000 units) or a fixed 2,000-unit dose 1, 2

• 4F-PCC is significantly less effective than andexanet for therapeutic FXaI concentrations 2
• Onset of action is within 10 minutes with hemostatic effect lasting approximately 8 hours 2

Common Pitfalls and Practical Considerations

Avoid These Mistakes

• Do not use andexanet for bleeding that responds to supportive measures alone—reserve it for life-threatening or uncontrolled hemorrhage 2
• Do not use it for isolated high drug levels without bleeding 2
• Do not use it for elective surgery that can be delayed to allow drug clearance 2
• Fresh frozen plasma is ineffective for DOAC reversal and should be avoided 2
• Recombinant factor VIIa should not be used as first-line reversal 2

Mandatory Post-Reversal Management

Resume therapeutic anticoagulation as soon as bleeding is controlled to mitigate the 10-18% thrombotic event risk 2, 3, 4

• The majority of thrombotic events occur in patients who do not restart anticoagulation 1, 2
• Consider bridging with prophylactic low-molecular-weight heparin if therapeutic anticoagulation must be delayed 4

Storage and Cost Barriers

• Refrigerated storage is required, which may limit availability in some settings 2
• Wholesale acquisition cost is $24,750 for low-dose and $49,500 for high-dose regimens 5

Edoxaban Reversal (Off-Label)

Use the high-dose regimen (800 mg bolus + 960 mg infusion) for edoxaban reversal, though this remains off-label 3, 6:

• In 28 evaluable patients with edoxaban-related bleeding, median anti-FXa activity decreased by 68.9% (95% CI 56.1-77.7%) 6
• 78.6% achieved excellent or good hemostasis at 12 hours (95% CI 59.0-91.7%) 6
• Thrombotic events occurred in 11.1% within 30 days, consistent with rates seen with apixaban/rivaroxaban 6

Bottom Line for Clinical Practice

Andexanet alfa provides rapid, substantial reversal of apixaban and rivaroxaban anticoagulation with demonstrated hemostatic efficacy in 67-80% of patients, but the 10% thrombotic event rate and lack of mortality benefit in ANNEXA-I demand careful patient selection and aggressive anticoagulation resumption strategies. 1, 2 The drug should be reserved for truly life-threatening bleeding where the immediate hemostatic benefit outweighs the prothrombotic risk, and clinicians must have a clear plan for anticoagulation resumption before administering it. 2, 3