Why this drug is interesting

Andexanet alfa is a recombinant form of human factor Xa modified so that it cannot cleave prothrombin. It is catalytically dead but structurally intact, so a direct factor Xa inhibitor binds it with the same avidity it would bind the real enzyme. Flood the circulation with decoy and the drug is sequestered away from the enzyme that matters.

It also inhibits tissue factor pathway inhibitor, which is a second, less-advertised mechanism and a partial explanation for both its efficacy and its thrombotic signal. The molecule is not simply a sponge; it is procoagulant in its own right.

What makes it worth studying is everything downstream of the pharmacology. The trials are positive on a surrogate and on a haemostatic endpoint. The label is specific and slightly counterintuitive. NICE recommends it for one bleeding site and not others. And the drug has a surgical interaction — unresponsiveness to heparin — that will catch out a team taking a reversed patient to theatre.

Indication and the funding position

The licence

"For adult patients treated with a direct factor Xa (FXa) inhibitor (apixaban or rivaroxaban) when reversal of anticoagulation is needed due to life-threatening or uncontrolled bleeding."¹ Restricted to hospital use only. It is a black triangle product, under additional monitoring.

NICE recommends it only for gastrointestinal bleeding

NICE TA697: "Andexanet alfa is recommended as an option for reversing anticoagulation from apixaban or rivaroxaban in adults with life-threatening or uncontrolled bleeding, only if: the bleed is in the gastrointestinal tract, and the company provides andexanet alfa according to the commercial arrangement."²

The committee's reasoning was explicitly economic rather than a judgement that the drug does not work: an indirect comparison suggested improved survival in gastrointestinal bleeding and intracranial haemorrhage but lower survival for bleeds elsewhere, with the comparison itself uncertain; and the cost-effectiveness estimates fell within acceptable limits only for gastrointestinal bleeding.

Intracranial haemorrhage remains outside NICE's recommendation. A separate appraisal (TA1029, January 2025) was terminated because the company did not make an evidence submission, taking the view that there was not currently enough evidence of cost-effectiveness in that population.³ So despite the positive ANNEXA-I result in intracranial haemorrhage, there is no NICE recommendation supporting its routine NHS use there. In practice this means access is governed by your local trust or network policy, and prothrombin complex concentrate remains the widely used alternative — check what yours says before you need it.

Dosing

Dose regimens (SmPC Table 1)
Regimen	Initial IV bolus	Continuous infusion	200 mg vials
Low dose	400 mg at ~30 mg/min (over 15 min)	4 mg/min for 120 min (480 mg)	5
High dose	800 mg at ~30 mg/min (over 30 min)	8 mg/min for 120 min (960 mg)	9

Which regimen — and the rule that summaries get wrong

Last dose taken	<8 hours before	≥8 hours before
Apixaban ≤5 mg	Low dose	Low dose
Apixaban >5 mg	High dose	Low dose
Rivaroxaban ≤10 mg	Low dose	Low dose
Rivaroxaban >10 mg	High dose	Low dose

The ≥8 hour column steps everything down to low dose, irrespective of how much was taken. This is easy to lose: several published summaries of this label state that a dose above the threshold requires the high-dose regimen regardless of timing, which is not what the table says. The regimen depends on both variables.

Administration

Reconstituted to 10 mg/mL and given without further dilution, via syringe pump or a polyolefin or PVC bag
A 0.2 or 0.22 micron inline low-protein-binding filter (PES or equivalent) is required before IV infusion
Restricted to hospital use
No dose adjustment for age, renal impairment or hepatic impairment
No data in children or adolescents — safety and efficacy not established

After the infusion

Restarting anticoagulation

The patient's underlying thrombotic risk has not gone away, and the drug has added to it. The label directs that antithrombotic therapy be re-initiated "as soon as medically indicated… if the patient is clinically stable and adequate haemostasis has been achieved," and states that on PK/PD modelling a normal degree of anticoagulation from apixaban, rivaroxaban or enoxaparin can be expected 4 hours after the end of the infusion.¹

Monitoring

"Treatment monitoring should be based mainly on clinical parameters indicative of appropriate response" — that is, has the bleeding stopped. Anti-FXa activity assays are unreliable after andexanet: commercial assays produce erroneously elevated anti-FXa levels, substantially overestimating residual anticoagulant effect. Do not chase the number.

Thrombosis and adverse effects

Adverse reactions

Common (≥1/100 to <1/10): ischaemic stroke, deep vein thrombosis, pulmonary embolism, myocardial infarction, pyrexia
Uncommon (≥1/1,000 to <1/100): transient ischaemic attack, cardiac arrest, arterial embolism, infusion-related reaction

Contraindications

Hypersensitivity to the active substance or excipients
Known allergic reaction to hamster proteins — the product is expressed in Chinese hamster ovary cells

The evidence

ANNEXA-4 — the licensing study

A single-arm, open-label study in patients with acute major bleeding on a factor Xa inhibitor. The label reports anti-FXa activity reductions of −93.3% for apixaban and −94.1% for rivaroxaban, with 79–80% of patients adjudicated as achieving good or excellent haemostasis.¹ There was no control group, which is the central limitation: 80% good haemostasis has no comparator, and most major bleeds stop with supportive care.

ANNEXA-I — the randomised trial

A randomised comparison against usual care in intracranial haemorrhage. The label reports 67.0% achieving effective haemostasis with andexanet alfa versus 53.1% with usual care (p=0.0032), with the thrombotic excess noted above.¹ The trial was stopped early for efficacy on this haemostatic endpoint.

Practical use in the ED

Establish the drug, the dose and the timing. Which agent, what strength, and when was the last tablet taken. All three change the answer, and "a blood thinner, yesterday" does not get you to a regimen.
Do the things that work regardless. Stop the anticoagulant, resuscitate, transfuse, correct calcium and temperature, get definitive haemostasis — endoscopy, interventional radiology, surgery. Reversal never substitutes for source control.
Check your local policy before assuming availability. NICE funds it for gastrointestinal bleeding. Many trusts hold small stocks or none, and intracranial use is a local network decision.
Involve haematology and pharmacy at the same moment. Nine vials do not appear quickly, and the unknown-dose scenario needs a specialist opinion rather than a protocol.
Hand over the two hazards explicitly: the patient is prothrombotic, and they will not respond normally to heparin. Both are easy to lose in a busy handover and both have caused harm.

Critical appraisal

The pivotal licensing evidence was uncontrolled. ANNEXA-4 had no comparator arm, and an 80% good-haemostasis rate in major bleeding is uninterpretable without one. The drug was licensed on a surrogate — anti-FXa reduction — plus a single-arm haemostasis rate, which is a low bar for a treatment with a 10% thrombotic event rate.
ANNEXA-I improved the evidence and did not settle the question. It is a genuine randomised trial with a positive haemostatic result, stopped early for efficacy — a design feature that tends to inflate effect estimates. It did not demonstrate a functional or mortality benefit, and it showed roughly double the thromboembolic events.
NICE's position is a cost-effectiveness judgement, not a clinical verdict — and clinicians frequently read it as the latter. The committee did not conclude that andexanet does not work in intracranial haemorrhage; it concluded the cost-effectiveness case was not made, and the subsequent ICH appraisal was terminated because the manufacturer declined to submit evidence. Those are different statements, and it is worth being precise about which one is being cited.
The comparison against prothrombin complex concentrate has never been made head-to-head in a trial that answers the question. PCC is cheaper, faster to give, familiar and available everywhere; it is also non-specific and thrombogenic in its own right. The indirect comparisons NICE reviewed were, in its own words, uncertain. That the obvious trial has not been done is the single biggest gap in this field.
The heparin unresponsiveness is under-appreciated in emergency practice. It is a label warning rather than a footnote, and it converts a reversal decision into a decision that constrains the next twelve hours of surgical management.

References

1
Ondexxya 200 mg powder for solution for infusion — Summary of Product Characteristics, AstraZeneca UK Limited. electronic Medicines Compendium. Sections 4.1–4.4, 4.8, 5.1–5.2, including Tables 1–3. Verified 23 Aug 2026.
2
NICE. Andexanet alfa for reversing anticoagulation from apixaban or rivaroxaban. Technology appraisal guidance TA697. nice.org.uk/guidance/ta697. Section 1 Recommendations. Verified 23 Aug 2026.
3
NICE. Andexanet alfa for reversing anticoagulation in people with intracranial haemorrhage (terminated appraisal). Technology appraisal guidance TA1029, 15 January 2025. nice.org.uk/guidance/ta1029
4
Connolly SJ, Crowther M, Eikelboom JW, et al. Full study report of andexanet alfa for bleeding associated with factor Xa inhibitors (ANNEXA-4). N Engl J Med 2019;380(14):1326–35. PubMed
5
Connolly SJ, Sharma M, Cohen AT, et al. Andexanet for factor Xa inhibitor-associated acute intracerebral hemorrhage (ANNEXA-I). N Engl J Med 2024;390(19):1745–55. PubMed
6
Andexanet alfa — monograph. BNF, NICE. bnf.nice.org.uk

Andexanet alfa

At a glance