Andexanet (ANDEXXA): A Blood Thinner Reversal Agent Whose Net Clinical Benefit Remains Unproven

In May 2018 the Food and Drug Administration (FDA) granted accelerated approval for the injectable drug andexanet (ANDEXXA). This drug is indicated to reverse life-threatening or uncontrolled bleeding caused by the use of two oral blood thinners approved for preventing blood clots and stroke: apixaban (ELIQUIS and generics) and rivaroxaban (XARELTO and generic).[1]

In the February 2021 issue of Worst Pills, Best Pills News, Public Citizen’s Health Research Group designated andexanet as Do Not Use because its clinical efficacy and safety had not been established.[2] Although andexanet supposedly reverses bleeding by binding and sequestering apixaban and rivaroxaban, it also can cause serious and life-threatening thromboembolic events (blood clots that break away and block blood vessels such as arteries [including those in the heart, causing a heart attack, or brain, causing an ischemic stroke]).[3] In fact, since the drug’s accelerated approval, the labeling of andexanet has contained a boxed warning — the most prominent safety warning required by the FDA — regarding these thromboembolic adverse effects (see the Text Box, below, for this warning).[4]

In November 2024 the FDA held a Cellular, Tissue, and Gene Therapies Advisory Committee meeting to discuss the results of an industry-sponsored post-marketing trial of andexanet, called ANNEXA-I.[5] This trial was required by the FDA in 2018 to assess the efficacy and safety of andexanet; however, the trial results were not published until 2024.[6] The advisory committee found that the efficacy results from the ANNEXA-I trial do not clearly show a clinically meaningful benefit for andexanet, reinforcing our Do Not Use designation of this drug. The committee also questioned andexanet’s safety profile because it increased the risk of thrombotic events (blood clots) in the trial.

Testifying in the open public hearing during the advisory committee meeting, Public Citizen’s Health Research Group urged the FDA to reject an application by the current manufacturer of andexanet (AstraZeneca) to convert the drug’s accelerated approval to full approval.[7] We made this recommendation because the ANNEXA-I trial failed to verify a potential net benefit for andexanet in patients with acute intracerebral hemorrhage (bleeding into the brain tissue) after receiving either apixaban or rivaroxaban.

The ANNEXA-I trial and the FDA decision

The ANNEXA-I trial enrolled adults who developed acute intracerebral bleeding within 15 hours of treatment with either apixaban or rivaroxaban. According to the FDA briefing material, data from 404 adult participants (204 who were randomized to receive andexanet and 200 who were randomized to receive usual care) were used for the primary efficacy analysis of the trial.[9]

In this trial, andexanet had a 12% advantage on a short-term primary endpoint of hemostatic efficacy over usual care; the observed rate of effective hemostasis was 65% in the andexanet group and 53% in the usual-care group. Measured 12 hours after randomization, this endpoint was a composite of three components: an expansion of the hematoma volume of 35% or less (as determined by imaging data), no use of rescue therapy, and minimal neurologic deterioration from baseline.

Although the 12% difference was statistically significant, its clinical meaningfulness is questionable because it was driven by hematoma expansion; the other two components of the primary endpoint differed only slightly between the two groups. Additionally, hematoma expansion 12 hours after randomization is an earlier time than recommended in National Heart, Lung, and Blood Institute guidelines. Because no imaging data were collected beyond 12 hours, delayed hematoma expansion (which is indicative of long-term benefit) is unknown. Moreover, because hematoma expansion was measured as a dichotomized response, it did not fully capture the treatment effect on this outcome. Importantly, the FDA had relayed some of these issues and others to AstraZeneca during the trial-design phase, but the drugmaker did not address them.

For safety, 15% of the andexanet-treated participants suffered thrombotic events, compared with only 7% of those who received usual care. This twofold increase associated with andexanet was statistically significant. It also was clinically important because it was driven by brain thrombosis in 9% of andexanet-treated participants, compared with 2% of those receiving usual care.

Although death rates were comparable between the two groups (28% for andexanet and 26% for usual care), deaths attributable to thrombotic events were more than twice as high in the andexanet-treated participants.

Therefore, the serious risks of andexanet outweighed its short-term benefit in the trial. As we discussed in our testimony, other important methodological issues not discussed in the FDA briefing material may suggest that certain aspects of the ANNEXA-I trial may have tipped the scales in favor of andexanet. For example, whereas both the safety analysis and sensitivity analyses for efficacy included 474 participants, the primary efficacy analysis excluded 70 of these participants. It is unclear whether this exclusion had favored andexanet’s primary efficacy results.

Additionally, given that 162 (79%) of andexanet-treated participants received a low dose of the drug, the findings may not be generalizable to those receiving a higher dose. Moreover, the fact that 11% of the usual-care participants received no treatment may have made the results worse than if all such participants had received at least some treatment.

In 2018, FDA clinical reviewers of the accelerated approval application for andexanet concluded that “the uncertainty regarding the clinical benefit [of andexanet], in combination with concerns about [its] safety, results in an unfavorable overall benefit-risk profile of [andexanet].”[10] However, the FDA director responsible for reviewing the drug application for accelerated approval downplayed this concern, arguing that risks are “mitigated” by drug-label warnings. Seven years later, there is still no acceptable net evidence to support andexanet.

Although the FDA informed AstraZeneca in December 2024 that it will not convert andexanet’s accelerated approval to full approval,[11] it is not clear whether the agency will request additional confirmatory studies for the drug or revoke its accelerated approval. So far, AstraZeneca has opted to keep this expensive drug on the U.S. market (priced in 2020 at $27,500 for a low dose in adults).[12]

What You Can Do

If you have a condition that requires you to use an oral blood thinner, discuss with your clinician the fact that there is currently no proven antidote for apixaban and rivaroxaban given the recent FDA decision not to grant andexanet full approval. Currently, we designate apixaban as Limited Use[13] and rivaroxaban as Do Not Use.[14] Although the blood thinner warfarin (JANTOVEN and generics) is cumbersome to use because, among other things, it requires frequent testing to monitor its levels in the blood, it has a proven and affordable antidote (vitamin K).

Unless you experience sudden, severe bleeding, do not stop your blood thinner medication suddenly without talking with your clinician because doing so can increase your risk of developing blood clots or stroke.

Seek immediate medical help if you fall or injure yourself (particularly if you hit your head) or if you experience any signs of bleeding, a blot clot or a stroke while taking any blood thinner.

References

[1] Food and Drug Administration. Accelerated approval letter for coagulation factor Xa (recombinant), inactivated-zhzo (BL125586/0). May 3, 2018. https://www.fda.gov/media/113285/download. Accessed March 6, 2025.

[2] Overview of the questionable drug andexanet (ANDEXXA). Worst Pills, Best Pills News. February 2021. https://www.worstpills.org/newsletters/view/1379. Accessed March 6, 2025.

[3] AstraZeneca. Label: andexanet alpha (Andexxa). March 2024. https://dailymed.nlm.nih.gov/dailymed/fda/fdaDrugXsl.cfm?setid=2d9d90a6-63e6-46ef-96ff-dd6519ae7b6c&type=display. Accessed March 6, 2025.

[4] Food and Drug Administration. Summary basis for regulatory action for ANDEXXA/coagulation factor Xa (recombinant), inactivated-zhzo. May 3, 2018. https://www.fda.gov/media/113954/download. Accessed March 6, 2025.

[5] Food and Drug Administration. Summary Minutes: 77th Cellular, Tissue, and Gene Therapies Advisory Committee meeting. November 21, 2024. https://www.fda.gov/media/185461/download. Accessed March 6, 2025.

[6] Connolly SJ, Sharma M, Cohen AT, et al. Andexanet for factor Xa inhibitor–associated acute intracerebral hemorrhage. N Engl J Med. 2024;390(19):1745-1755.

[7] Public Citizen. Testimony before the FDA’s Cellular, Tissue, and Gene Therapies Advisory Committee regarding andexanet (Andexxa). November 21, 2024. https://www.citizen.org/wp-content/uploads/2704.pdf. Accessed March 6, 2025.

[8] AstraZeneca. Label: andexanet alpha (Andexxa). March 2024. https://dailymed.nlm.nih.gov/dailymed/fda/fdaDrugXsl.cfm?setid=2d9d90a6-63e6-46ef-96ff-dd6519ae7b6c&type=display. Accessed March 6, 2025.

[9] Food and Drug Administration. FDA briefing document, supplemental application sBLA# 125586/546 for Andexxa (coagulation factor Xa [recombinant], inactivated-zhzo), Cellular, Tissue, and Gene Therapies Advisory Committee Meeting. November 21, 2024. https://www.fda.gov/media/183674/download. Accessed March 6, 2025.

[10] Food and Drug Administration. Summary basis for regulatory action for ANDEXXA/coagulation factor Xa (recombinant), inactivated-zhzo. May 3, 2018. https://www.fda.gov/media/113954/download. Accessed March 6, 2025.

[11] Dunleavy K. FDA rejects J&J's subcutaneous Rybrevant filing and AZ's full approval bid for Andexxa. December 17, 2024. https://www.fiercepharma.com/pharma/fda-rejects-johnson-johnsons-subcutaneous-rybrevant-astrazenecas-andexxa. Accessed March 6, 2025.

[12] Mahan CE. Reply to “key points to consider when evaluating Andexxa for formulary addition.” Neurocrit Care. 2020;33(1):323-326.

[13] The blood thinner apixaban (ELIQUIS): An Update. Worst Pills, Best Pills News. April 2024. https://www.worstpills.org/newsletters/view/1589. Accessed March 36, 2025.

[14] Overview of the blood thinner rivaroxaban (XARELTO): An update. Worst Pills, Best Pills News. May 2021. https://www.worstpills.org/newsletters/view/1397. Accessed March 36, 2025.