JC: Reversal of DOACs with Andexanet Alfa. St.Emlyn’s

123We are in a new age of anticoagulation, Warfarin is becoming a drug of the past. To the benefit of our patients we are no longer dragging them into hospital once a week to check their INR. The so-called “DOACs” (Direct Oral Anticoagulants – Apixaban, Rivaroxaban) are being used to anti-coagulate patients with PEs, DVTs and AF. We have covered this here on St.Emlyn’s before: click here for DOAC and anticoag blogs from St.Emlyn’s

However this can be a scary prospect to the emergency physician. What do we do with the bleeding patient on these DOACs? How do we reverse them safely and effectively? This week we have a trial in the NEJM4 tackling this very question involving Andexanet Alfa, studying its effects on the reversal of DOACs in the bleeding patient. Click on the picture below to read the abstract on the NEJM site (and the full paper if you can).

So what is Andexanet Alfa?

If you are anything like me, you may shudder at the thought of having to recall the clotting cascade!

When I was at medical school I learnt about a clotting cascade, involving an extrinsic and intrinsic pathway. Things have moved on a bit since then; clotting is described as a cell based model with four phases: initiation, amplification, propagation and stabilisation. Factor Xa is involved in the initiation stage of the clotting cascade. For a closer look please see the images bellow.

Vessel wall rupture 





Tear in the vessel wall causes a Tissue Factor (TF) bearing cell to bind with activated factor VIIa, producing some thrombin and  factor IXa 1–3



amplificationThrombin causes the activation of platelets, producing clotting factors Va, VIIIa and XIa 1–3



These activated clotting factors and platelets produce factor Xa. Xa and Va combine to produce large quantities of thrombin, which leads to the production of fibrin 1–3

Andexanet Alfa (andexanet) is a factor Xa protein that has been modified to be inactive so it is unable to activate prothrombin and have a high affinity for factor Xa inhibitors.  This means it binds to small molecule factor Xa inhibitors such as Apixaban and Rivaroxaban

Back to the study: What did they do?

This an interim report of a study that is still actively recruiting, which is looking at the reversal of anticoagulation in patients who were taking Apixaban, Rivaroxaban, Enoxaparin or Edoxaban (i.e. a Factor Xa inhibitor) who presented to hospital with acute major bleeding. Major bleeding was defined as any of the following: bleeding patients who are haemodynamically unstable, or with a fall in Hb > 2g/dl, a bleed that was predicted to cause Hb to drop below 8 g/dl, or in which Hb was less than 8g/dl, symptomatic bleeding in a critical organ such as pericardial or intracranial bleed that has been confirmed by a CT scan.

All patients who met the inclusion criteria were given a predetermined bolus dose of Andexanet Alfa followed by a 2 hour infusion of the drug.

During this time patients were assessed multiple times with follow up, and up to 30 days following infusion. Bloods were taken to measure anti-factor Xa levels as well as levels of unbound factor Xa inhibitor in the plasma.

The primary end point was achievement of haemostasis 24 hours from the start of treatment.  Haemostasis was classified as excellent or good. For intracranial haemorrhage this was defined by percentage increase in haemorrhage volume. For GI and other non-visible bleeds, results were defined by percentage decrease from baseline in Hb and haematocrit. In patients with visible bleeding, if bleeding stopped within one hour of infusion starting this was considered excellent and with 4 hours was considered good.

Secondary endpoints were considered in terms of percentage change from baseline anti factor Xa levels.

Who was included?

So far a total of 67 adults having an acute major bleed requiring reversal of anti-coagulation have been recruited. Following recruitment, 20 patients were excluded from the trial, leaving 47 in this data analysis.

Who was excluded?

The usual suspects were excluded (children, pregnant women etc) as well as patients whose bleeding could not be stopped solely by reversal of the anticoagulant (ie traumatic bleeds or vessel rupture), patients who had recently (within 14 days) been given drugs that would affect coagulation (i.e. prothrombin complex, clotting factors and platelets), patients who had had a thrombotic event in the last 2 weeks and patients with acute coagulopathies such as DIC and septic shock.

Population characteristics 

The majority of patients in the trial were anticoagulated due to atrial fibrillation (68%), the remaining patients had venous thromboembolism (26%) or mixed pathologies.

A total of 26 patients were taking 20mg (IQR 15-20) of Rivaroxaban daily. Apixaban was the anti-coagualtion of choice for the other 20 patients, who were taking a median daily dose of 5mg (IQR 5-10). Prior to treatment, baseline antifactor Xa activity was measured; 297 (SD 1714.0) ng/ml in the Rivaroxaban patients and 174.5 (SD 97) in the Apixaban group. In both of these groups there is a large standard deviation. Considering we prescribe these anticoagulants as a “one size fits all” therapy, this does make me wonder if this is actually the case.

GI bleeding was the commonest cause for admission (25 of 47 patients), with higher incidences in the Rivaroxaban group (64% vs 32%). When looking at the intracranial bleeds the commonest site was intracerebral (60%) followed by subdural (35%), with higher incidence in the Apixaban group (60% vs 40%).

What did they find?

At first look the clinical and pharmacodynamic outcomes look promising. 31 patients were assessed as having excellent haemostasis and 6 patients labelled good haemostasis at 12 hours post infusion. So that’s good or excellent haemostasis in 79% (but the 95% confidence intervals are wide at 64-89%, thus it’s not a very precise finding).

When looking at Rivaroxaban vs Apixaban, there were high rates of success in both with little difference between them. Good or excellent haemostasis was achieved in 81% (21/26) of the Rivaroxaban group with a 95% CI of 61-93%, and in 75% of the Apixaban group with a 95% CI of 51-91%. In the 8 patients assessed as having poor or no haemostasis, 5 were taking Rivaroxaban and 3 Apixaban. With such small numbers we are again hampered by wide confidence intervals.

Site of bleeding also did not seem to make a difference. Good or excellent haemostasis was achieved in 80% of GI bleeds (95% CI 64-94) and 80% of intracranial bleeds (95% CI 56-94).

Laboratory values of anti factor Xa levels backed up these clinical findings. Following a bolus of Andexanet, anti-factor Xa levels fell 89% from baseline in the Rivaroxaban group (95% CI 58-94%) and 93% from baseline in the Apixaban group (95% CI 87-94%).

Is this all good news?

This study aims to recruit 167 patients in order to power it sufficiently, so there is a long way to go before we can get very excited but these initial results are promising.

There are, however, some limitations; while primary end points have mainly remained defined by numerical laboratory values some are still observer-dependent, which might allows a element of bias to creep in as there is no blinding in this study. The interim results have very wide confidence intervals although this may change when the study is sufficiently powered.

Editor –  Several of the St.Emlyn’s team are uncertain why this has been published now, in a journal with global prestige, yet whilst the treatment effect is still uncertain. This is an interim analysis but there is no doubt that a publication in the NEJM will influence the adoption of this therapy. Indeed there is a precedent for this, again involving a DOAC reversal agent, again in the NEJM, and again an interim analysis: Idarucizumab for the reversal of Dabigatran appears to have followed a similar research and publication process which is unusual. I’m not sure I’d ever get an interim analysis published in the NEJM.

There was a safety group in this trial, in which 18 out of 67 patients died, giving a 27% 30-day mortality. 12 of these patients had a thrombotic event within 30 days of receiving Andexanet. The paper doesn’t give the number of thrombotic events in the patients who did not die. Death is obviously the most serious adverse event, however I would like to see more data on adverse affects in order to look at number need to harm against number needed to treat.

I am looking forward to seeing the results of the trial once completed. It’s definitely worth asking why an interim report has been published. They do not have enough data to show the safe reversal of DOACs on the basis of this analysis. I hope the trial continues to successfully recruit as we do need a definitive answer and reversal agent for these drugs, and as an EM physician the prospect of having a safe reversal agent for patients who are bleeding and taking DOACs is exciting. However, further work in this area is clearly needed before we can be assured that this is the magic bullet of reversal.





A cell-based model of coagulation and the role of factor VIIa. Anaesthesia UK. http://www.frca.co.uk/article.aspx?articleid=100097. Accessed December 9, 2016. [Source]
A new model of Hemostasis. Heparin Science . http://www.heparinscience.com/model_hemostasis.html. Accessed December 9, 2016.
Call based coagualtion . reddy medical communications. http://reddymed.com/cbc/cbc.html. Accessed December 9, 2016.
Connolly SJ, Milling TJ Jr, Eikelboom JW, et al. Andexanet Alfa for Acute Major Bleeding Associated with Factor Xa Inhibitors. New England Journal of Medicine. 2016;375(12):1131-1141. doi: 10.1056/nejmoa1607887


  1. H

    Excellent post Laura – thanks

  2. Pingback: JC: Reversal of DOACs with Andexanet Alfa. St.Emlyn’s – Global Intensive Care

  3. Gareth

    Great post and great editorial comments.


Thanks so much for following. Viva la #FOAMed

Translate »