Goodbye to Andexanet Alfa (AndexxaⓇ)

Andexanet alfa (Andexxa) made its quiet exit from the U.S. market just a few weeks ago.  It's an interesting story so in case you missed it, I'll recap what andexanet alfa is, or rather was, and what happened that led to its removal.  Being pulled from the U.S. market may have come to a surprise for some considering its main study (ANNEXA-I) was stopped early for successfully meeting its main efficacy outcome. While having a trial stopped early for superior efficacy results sounds good on the surface, it can have negative ramifications on other aspects of the clinical inquiry (explained below).

Andexanet alfa was approved for patients treated with rivaroxaban (Xarelto) or apixaban (Eliquis) when reversal of anticoagulation was needed for life-threatening or uncontrolled bleeding.

 

Background studies and initial approval 

Andexanet alfa was initially granted "Accelerated Approval" from the U.S. Food and Drug Administration (FDA) in 2018 following 3 studies:  

• ANNEXA-A and ANNEXA-R were randomized trials in healthy individuals who were given apixaban and rivaroxaban, respectively.  These patients were then treated with andexanet alfa and had their anti-factor Xa concentrations assessed.  Andexanet alfa successfully corrected the anti-factor Xa elevations in both studies.  
• Then, in the single arm study in patients with major bleeds (ANNEXA-4), andexanet alfa had the same effect on hemostatic laboratory values plus seemed to have other hemostatic benefits.  For a more thorough discussion about these studies, see my previously published narrative review article here.

 

What is "Accelerated Approval" from the U.S. FDA?

Accelerated Approval is a program from the U.S. FDA that allows earlier approval of drugs that are meant to treat an unmet need.  The approvals can be based on a surrogate marker (such as a laboratory or imaging result) for an important endpoint, rather than the clinical endpoint itself.  A surrogate endpoint, in general, might be appropriate to first assess if we believe it represents or estimates the more meaningful clinical events that will happen later (and surrogates are often easier or more practical to detect and we don't have to wait as long).  

Importantly, Accelerated Approval requires a follow up study after the medication is on the market to assess the actual clinical endpoint the medication is hypothesized to achieve and the net risk-benefit relationship.  If the follow up study isn't done or the evidence doesn't satisfy the FDA, there is a mechanism to withdraw it from the market.

 

Follow up clinical trial

Following the FDA's Accelerated Approval, the manufacturer was required to do a follow up randomized controlled trial in patients with intracerebral hemorrhage (following rivaroxaban or apixaban use) to ensure there was net clinical benefit.  This study was the previously mentioned ANNEXA-I which revealed that andexanet alfa did, in fact, have superior hemostatic efficacy compared to usual care (specifically, this was a combination of hematoma volume expansion, NIHSS score change, and the need for rescue therapy).  The downside was that more patients experienced thrombotic events after receiving andexanet alfa versus usual care (10.3% vs. 5.6% in the NEJM article).  The trial was stopped early by the data and safety monitoring board for reaching the efficacy endpoint.

At this point, AstraZeneca, andexanet alfa's manufacturer, applied for a supplemental Biologics Licensing Application for Andexxa, hoping to fulfill the FDA requirement and be converted to a traditional approval (not just Accelerated Approval).  During the FDA Advisory Committee meeting to decide andexanet alfa's fate, it was revealed that the FDA's review of the data identified additional patients with thrombotic events, bringing their thromboembolic event result to 14.6% versus 6.9% in the andexanet alfa and usual care arms, respectively.  (Why the difference in event rates? While the trial authors of ANNEXA-I had a prespecified charter on how they defined the events, the FDA was also sent the data and ended up identifying additional events.)  Besides the increased thromboembolic rate with andexanet alfa, the committee members had additional concerns, discussed in this next section.

 

How stopping the trial early could have hurt andexanet alfa's chances (or maybe revealed something worse)

Once again, andexanet alfa was (1) proven to have improved hemostatic effects as far as anti-factor Xa concentrations but (2) worse effects as far as thromboembolic events.  Importantly, in intracranial hemorrhage studies, the outcomes most commonly assessed and also most highly valued are not these first two endpoints.  Those would be:

1. mortality (93% of studies) and  
2. functional outcomes (85% of studies - usually modified Rankin Scores) 

Also, these outcomes are typically studied at more distant endpoints than the ANNEXA-I study analyzed (the vast majority of literature analyzes these important outcomes at 90 days, 180 days, or longer).  The ANNEXA-I study only reached 30 day outcomes and found no difference in either mortality or modified Rankin scores.  If the study were followed for the longer duration, perhaps we would have answers as to the effect of andexanet alfa on these outcomes.  

Unfortunately for andexanet alfa, the 30 day outcomes suggested the andexanet alfa arm might have done worse.  For example, in patients that had a baseline modified Rankin Score of 0-3 (meaning they could walk independently) were more likely to progress to a score of 4-6 (meaning they could walk only with assistance or died) if they received andexanet alfa compared to usual care (68% vs. 61%, respectively).  This means that the early signal from the ANNEXA-I trial revealed the more highly functional patients at baseline ended up with worse functional independence at 30 days.

Here's the 6 hour recording of the Advisory Committee if you're interested in the nitty gritty details. 

 

Take away 

So regardless of how thromboembolic events were adjudicated, there were significantly more thromboembolic events in the andexanet alfa group and stopping the trial early prevented the true analysis of more meaningful outcomes such as mortality or functional outcomes at the typically appreciated time points (there was no difference in these at the limited 30 day assessment other than a signal that certain andexanet alfa patients actually did worse).  This all led the Advisory Committee not to award full approval. They decided the company failed to prove the hypothesized benefit that was based on the surrogate endpoints prior to marketing and that ultimately, the risks outweigh the benefits.  Then, on December 18, 2025, the FDA communicated that AstraZeneca was withdrawing the approval application and would discontinue manufacturing and sales in the U.S. by December 22, 2025.

References - all hyperlinked above

 

image by Sharon Sinclair (no changes were made)