ANNEXA-4: Andexanet Alfa for FXa-Inhibitor-Associated Major Bleeding (Single-Arm Cohort)
In adults with acute major bleeding (64% intracranial, 26% gastrointestinal) within 18 hours of an FXa-inhibitor dose (apixaban, rivaroxaban, edoxaban, enoxaparin), does andexanet alfa per the FDA-label algorithm reduce anti-FXa activity and achieve hemostatic efficacy at 12 hours? Single-arm prospective cohort that underwrote FDA accelerated approval in May 2018 — no comparator, no efficacy inference in the strict sense.
Connolly SJ, et al. for the ANNEXA-4 Investigators (NEJM 2019;380:1326-1335) · doi:10.1056/NEJMoa1814051 · 352 patients
Population
Included
- Adults with acute major bleeding requiring urgent reversal
- Most recent dose of apixaban, rivaroxaban, edoxaban, or enoxaparin within 18 hours before andexanet bolus
- For intracranial bleeding: acceptable functional status (modified Rankin Scale at most 3 before the event)
- Hemodynamic and clinical stability sufficient to receive treatment
- At least 18 years of age
- Presented with acute major bleeding
- Had received within 18 hours one of the following: apixaban, rivaroxaban, or edoxaban at any dose, or enoxaparin at a dose of at least 1 mg per kilogram of body weight per day
Acute major bleeding (defined as one or more of the following)
- Potentially life-threatening bleeding with signs or symptoms of hemodynamic compromise (e.g., severe hypotension, poor skin perfusion, mental confusion, or low cardiac output that could not otherwise be explained)
- Bleeding associated with a decrease in the hemoglobin level of at least 2 g per deciliter (or a hemoglobin level of <=8 g per deciliter if no baseline hemoglobin level was available)
- Bleeding in a critical area or organ (e.g., retroperitoneal, intraarticular, pericardial, epidural, or intracranial bleeding, or intramuscular bleeding with compartment syndrome)
Excluded
- Expected survival less than 1 month
- Planned major surgery within 12 hours
- Recent (within 2 weeks) thrombotic event
- Severe coagulopathy unrelated to FXa-inhibitor use
- Planned surgery within 12 hours after andexanet treatment (with the exception of minimally invasive operations or procedures)
- Intracranial hemorrhage in a patient with a score of less than 7 on the Glasgow Coma Scale or an estimated hematoma volume of more than 60 cc
- Expected survival of less than 1 month
- The occurrence of a thrombotic event within 2 weeks before enrollment
- Use of any of the following agents within the previous 7 days: vitamin K antagonist, dabigatran, prothrombin complex concentrate, recombinant factor VIIa, whole blood, or plasma
Source: Connolly SJ et al. (ANNEXA-4), NEJM 2019;380(14):1326-1335; PMC6699827· Retrieved 2026-06-09
Primary Outcome (descriptive, single arm)
Coprimary: percent change in anti-FXa activity from baseline to nadir; rate of excellent or good hemostatic efficacy at 12 h after infusion end
Excellent or good hemostasis at 12 h
Single-arm observational design — no efficacy inference possible. Anti-FXa fell a median 92% (95% CI 91–93% apixaban; 88–94% rivaroxaban). Hemostatic efficacy 204/249 (82%) at 12 h. 30-day mortality 14%; thrombotic events 10% — neither attributable causally to andexanet. The randomized ICH-specific confirmation is ANNEXA-I (NEJM 2024), which quantified the ischemic-stroke trade-off (NNH ~20) that this single-arm dataset could only describe.
Study Arms
- Agent
- Andexanet alfa (recombinant modified factor Xa decoy protein)
- Dose
- Low dose: 400 mg bolus + 480 mg infusion. High dose: 800 mg bolus + 960 mg infusion. Regimen selected by FXa inhibitor and timing of last dose.
- Route
- IV
- Frequency
- Bolus over 15-30 minutes followed by a 2-hour continuous infusion
- Duration
- Bolus plus 2-hour infusion
- Co-interventions
- No randomized comparator (single-arm prospective cohort). Standard supportive care for major bleeding.
Trial protocol dosing verbatim from Connolly SJ et al., NEJM 2019: the low-dose regimen (400 mg bolus over 15 minutes, 480 mg infusion) was used when apixaban or rivaroxaban had been taken more than 7 hours before; the high-dose regimen (800 mg bolus over 30 minutes, 960 mg infusion) was used for enoxaparin, edoxaban, or rivaroxaban taken 7 hours or less before or at an unknown time. Each vial also delivers mannitol (osmotic load relevant in ICH). The FDA-label algorithm expresses dose selection by last-dose amount and timing; see this entry applicability.doseSpecific. No control arm: efficacy cannot be inferred. Median anti-FXa activity fell ~92%; excellent or good hemostatic efficacy at 12 h in 204/249 (82%).
Safety
30-day mortality (single arm)
14%
0%
30-day mortality 49 of 352 (14%) in the andexanet cohort. No comparator arm; this is a descriptive event rate in an elderly population with major bleeding (mean age 77).
Any thrombotic event within 30 days (single arm)
10%
0%
34 of 352 (10%) patients had a thrombotic event within 30 days. Causality cannot be inferred from a single-arm design; subsequent randomized ANNEXA-I (Connolly NEJM 2024) quantified the ischemic-stroke excess vs usual care.
Trial Design
Type
- Single-arm, open-label, prospective cohort study (no comparator arm)
- Conducted to support FDA accelerated approval (granted May 2018)
- 63 centers across North America and Europe
- Adjudicated hemostatic efficacy (excellent/good/poor/none) at 12 h after infusion end
Timeline
Enrolled April 2015 to May 2018; 30-day follow-up
N
352
Enrollment
352 patients enrolled (254 in efficacy population) at 63 centers in North America and Europe. Enrolled April 2015 to May 2018. Single-arm, open-label, prospective cohort — no randomization, no comparator. Adjudicated hemostatic efficacy (excellent/good/poor/none) at 12 h after infusion end. 30-day follow-up. Supported FDA accelerated approval (granted May 2018). Sponsor: Portola Pharmaceuticals (now Alexion AstraZeneca Rare Disease). Published NEJM 2019.
Bedside Pearl
For acute major bleeding on apixaban, rivaroxaban, edoxaban, or enoxaparin within 18 h of last dose, andexanet alfa achieved 92% anti-FXa reduction and 82% excellent/good hemostasis at 12 h in 352 patients (ANNEXA-4). AHA/ASA 2022 Class IIa, Level B-NR for FXa-inhibitor-associated ICH. Single-arm cohort; efficacy was randomized-confirmed in ICH only by the later ANNEXA-I (NEJM 2024), which also quantified the ischemic-stroke trade-off.