Factor XI inhibitors and atrial fibrillation: imminent breakthrough or false start?

Abstract

Atrial fibrillation (AF) is a common cardiac arrhythmia associated with a high risk of thrombo-embolic events, such as ischaemic stroke and systemic embolism, which require anticoagulant treatment. Vitamin K antagonists and direct oral anticoagulants represent the current therapeutic standards, but they are limited by the risk of bleeding. In this scenario, factor XI (FXI) inhibitors are emerging as a new therapeutic option, potentially capable of reducing bleeding risk while maintaining antithrombotic efficacy. Molecules such as abelacimab, asundexian, and milvexian are under investigation for the prevention of thrombo-embolic events in patients with AF. Although preliminary data on these compounds suggest a favourable safety profile, the results regarding efficacy do not yet appear convincing. The phase 2 AZALEA-TIMI 71 trial was prematurely terminated after demonstrating a clear reduction in the incidence of major bleeding with abelacimab compared to rivaroxaban, whereas the phase 3 OCEANIC-AF study on asundexian was stopped due to inferiority compared to apixaban. Ongoing trials, such as LILAC-TIMI 76 and LIBREXIA-AF, are crucial to confirm the efficacy and safety of this therapeutic class. While FXI inhibitors represent a potential breakthrough in the treatment of AF, further data are needed to determine their definitive role in clinical practice.

Introduction

Atrial fibrillation (AF) is one of the most common cardiac arrhythmias, affecting over 33 million individuals worldwide.¹ Its prevalence increases with age, reaching ∼10% among those over 80, and is expected to double in the coming decades, driven by an ageing population, rising comorbidities, and advancements in diagnostics. These trends carry significant implications for healthcare systems and substantial socioeconomic repercussions.^1,2

Atrial fibrillation is associated with a wide range of severe adverse events, as well as increased mortality.¹ Among the most significant and frequent complications of AF are ischaemic stroke and systemic embolism (SE).¹ To assess the risk of these events in patients with AF, the CHA₂DS₂-VA score is available. This tool considers factors such as age, chronic heart failure, hypertension, diabetes, vascular disease, and a prior history of stroke or transient ischaemic attack.¹

Oral anticoagulant (OAC) therapy is the standard of care for preventing stroke and SE in patients with AF.¹ Currently available pharmacological options include vitamin K antagonists (VKAs) and direct oral anticoagulants (DOACs).¹ Although all DOACs have demonstrated at least non-inferior efficacy compared to VKAs and are associated with a lower risk of intracranial haemorrhage, both drug classes have significant limitations that make anticoagulant treatment challenging or even impossible for certain categories of patients at high risk of bleeding. For this reason, new anticoagulant molecules are currently under development, aiming to overcome these limitations.

Current recommendations on anticoagulant therapy in atrial fibrillation

The 2024 guidelines of the European Society of Cardiology on the management of AF recommend the use of the CHA₂DS₂-VA score as a tool to assess the eligibility of patients with this arrhythmia for initiating OAC therapy.² Specifically, anticoagulation is recommended for patients with a score ≥ 2 [Class of Recommendation (COR) I, Level of Evidence (LOE) C], while in patients with a score of 1, anticoagulant treatment should be considered (COR IIa, LOE C).² These recommendations apply regardless of the type of AF (paroxysmal, persistent, or permanent).²

The choice of OAC preferably favours the use of DOACs (COR I, LOE A).² However, there are specific situations in which the use of VKAs is exclusively indicated. Specifically, VKAs are recommended when AF is accompanied by the presence of a mechanical heart valve or moderate-to-severe mitral stenosis, due to the superior risk-benefit profile demonstrated in these specific clinical settings (COR I, LOE A).² Conversely, in patients with AF and a bioprosthetic valve, the use of DOACs is not only recommended but also preferred over VKAs.²

To optimize the balance between reducing thrombo-embolic risk and minimizing bleeding risk, it is essential to maintain International Normalized Ratio values within the therapeutic range of 2.0–3.0 in patients receiving VKAs (COR I, LOE B).² For patients on DOAC therapy, standard dosing is recommended, except in specific clinical conditions that justify dose reduction, such as advanced stages of chronic kidney disease.² Inappropriate dose reduction is discouraged, as it is associated with an increased ischaemic risk without a corresponding significant benefit in bleeding risk reduction (COR III, LOE B).²

Finally, the use of antiplatelet drugs is not recommended as an alternative to OACs for the prevention of ischaemic stroke and SE in patients with AF (COR III, LOE A).²

Unmet therapeutic needs of current anticoagulant strategies in atrial fibrillation

The efficacy of VKAs and DOACs in preventing ischaemic stroke and SE is well established.² The introduction of DOACs marked a significant therapeutic advancement due to their more favourable risk-benefit profile, more predictable pharmacokinetics, fewer drug-drug interactions, and greater ease of use compared to VKAs, thanks to the lack of routine laboratory monitoring requirements.^2,3 These benefits have also been confirmed in patients with AF who are ineligible for treatment with VKAs. In this context, the AVERROES trial (Apixaban Versus Acetylsalicylic Acid to Prevent Stroke in Atrial Fibrillation Patients Who Have Failed or Are Unsuitable for Vitamin K Antagonist Treatment) demonstrated that apixaban, compared to aspirin, offers greater efficacy in preventing ischaemic cerebrovascular events without a significant increase in the risk of major bleeding or intracranial haemorrhage.⁴ However, despite DOACs being associated with a generally lower bleeding risk—except for gastrointestinal bleeding—the incidence of severe bleeding events remains considerable.³ Randomized clinical trials report annual major bleeding rates of 2–3%.⁵ Key predisposing factors include age, frailty, and comorbidities, particularly advanced stages of hepatic or renal disease, which can significantly alter the pharmacokinetics of OACs, compromising their safety. Furthermore, as demonstrated in a post hoc analysis of the ARISTOTLE trial (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation), major bleeding events are associated with a significantly increased risk of death, ischaemic stroke, and myocardial infarction within 30 days of the index event.⁶ The explanation for these adverse outcomes lies, at least in part, in the interruption of anticoagulant therapy that often follows bleeding episodes.⁵

Several approaches have been developed to reduce the incidence of adverse events related to OAC use. Among these, the integration of predictive tools, such as the HAS-BLED score, has enabled personalization of therapy based on bleeding risk. Similarly, the use of low-dose DOAC regimens has shown favourable results in selected patient subgroups, as demonstrated, for example, by the use of edoxaban 15 mg once a day in elderly subjects in the ELDERCARE-AF trial (Edoxaban Low-Dose for Elder Care in Atrial Fibrillation).^7,8 In addition, the development of reversal agents such as idarucizumab and andexanet has provided effective antidotes, although their clinical implementation remains currently limited, mainly due to high associated costs.^9,10

Despite these efforts, ∼66% of patients worldwide do not receive OAC therapy despite clear guideline-based indications, primarily due to physicians’ concerns about bleeding risk.¹¹ Adding to this, the proportion of AF patients with concomitant contraindications to OACs is estimated to be around 12% in some cohorts.¹² Although percutaneous left atrial appendage occlusion is an available option for such patients, the need for post-procedural antithrombotic therapy to prevent device-related thrombosis risks undermining the expected benefits.

This discussion highlights the urgent need to develop new anticoagulant drugs with an improved safety profile, capable of reducing bleeding risk while maintaining the current standards of clinical efficacy.

Factor XI inhibitors as a new therapeutic frontier

The role of factor XI in haemostasis and thrombosis

The traditional concept of the coagulation cascade, originally described as a process stemming from two parallel pathways—the extrinsic pathway and the intrinsic pathway—converging into the common pathway, has been progressively replaced by the so-called cell-based model.¹³ This updated model integrates the two pathways into an interdependent sequence, emphasizing their functional interconnection (Figure 1). According to this model, activation of the extrinsic pathway serves as the primum movens in haemostasis and thrombosis, resulting from the interaction between tissue factor (TF) and factor VII—an event typically occurring following vascular wall injury.¹³

The resulting TF/FVII activated complex (FVIIa) is responsible, in the initial phase of the process, for producing a limited amount of thrombin. Thrombin, in turn, activates cofactors (factors V and VIII), while the TF/FVIIa complex leads to the activation of factor IX.¹³ This process amplifies thrombin production, whereas at this stage, the role of factor XI (FXI) is limited to the stabilization of the haemostatic plug.¹³

In the advanced phase of thrombotic events, FXI further activates factors IX, VIII, X, and V through a positive feedback mechanism, thereby promoting thrombus expansion and growth.¹³ Thrombosis is further amplified by the presence of natural polyanions within the thrombus structure, which in turn promote the activation of FXI.^13,14 When the coagulation cascade is triggered by contact mechanisms—such as blood interaction with artificial surfaces—FXI activation is primarily mediated by factor XII.^13,14

Rationale for factor XI inhibition

FXI is a zymogen primarily synthesized in the liver.¹⁵ Its activation into a functional serine protease occurs mainly through thrombin, activated factor XII, and natural polyanions.¹⁵ FXI deficiency is responsible for haemophilia C, or Rosenthal syndrome, which, as documented by numerous experimental and epidemiological studies conducted in humans and animal models, is typically associated with moderate bleeding, particularly following trauma, surgical procedures, or childbirth.^14,15 In contrast to haemophilia C, the bleeding observed in haemophilia A and B, caused by deficiencies in factors VIII and IX, respectively, generally occurs spontaneously and is more severe.¹⁵

On the other hand, while FXI deficiency is associated with a reduction in thrombotic risk, elevated plasma concentrations of this factor may promote ischaemic events, such as stroke or venous thromboembolism (VTE), whereas its role in myocardial infarction remains less clear.^14,15 This effect is explained by the role of FXI in thrombus amplification and growth mechanisms.¹⁵

Understanding the role of FXI in haemostatic processes has encouraged the development of innovative pharmacological strategies aimed at dissociating haemostasis from thrombosis through FXI inhibition.¹⁵ The goal is to develop new classes of drugs capable of effectively preventing ischaemic events while minimizing the risk of bleeding complications.

Clinical development of factor XI inhibitors

Molecules developed to inhibit FXI and its activated form, currently under clinical investigation, include small synthetic molecules, monoclonal antibodies, and antisense oligonucleotides, while natural peptides, aptamers, and siRNA have so far been studied only in preclinical settings.^13,16 The main clinical areas of application for FXI inhibitors include postoperative VTE prevention following knee arthroplasty, end-stage chronic kidney disease, non-cardioembolic ischaemic stroke, cancer, acute coronary syndromes, and AF.^13,16 Among these, VTE prevention represents one of the most promising scenarios for FXI inhibitors, as evidenced in phase 2 clinical studies, particularly when compared to enoxaparin.^13,16 Conversely, studies conducted in patients with non-cardioembolic ischaemic stroke have not yet demonstrated significant therapeutic benefits, though limited by the sample sizes typical of phase 2 trials. Similarly, the efficacy in acute coronary events remains a debated topic, despite the substantial, dose-dependent inhibition of FXI levels observed with the tested agents.^13,14 Anyway, FXI inhibitors have shown a favourable safety profile across all clinical settings examined.¹³

The FXI-targeting anticoagulants currently under investigation for AF include a monoclonal antibody, abelacimab, and two small synthetic molecules, asundexian and milvexian. The pharmacological characteristics of these compounds are summarized in Tale 1.

Randomized controlled trials on factor XI/FXIa inhibitors in atrial fibrillation

Completed studies

Currently available data on abelacimab come exclusively from phase 2 studies. Among these, ANT-004 (A Dose-range Finding Study of MAA868 in Patients With Atrial Fibrillation) was a small, randomized, double-blind, placebo-controlled study that evaluated the pharmacokinetic and pharmacodynamic characteristics of this compound, as well as its safety profile.¹⁷ Conducted from December 2019 to September 2020, the trial involved 18 patients who were randomized to receive a monthly subcutaneous dose of abelacimab (120 mg or 180 mg) or placebo for three months. The study was halted in September 2020, as decided by the Sponsor, due to operational challenges related to the SARS-CoV-2 pandemic. Participants, aged 18–85 years, diagnosed with AF or atrial flutter, had a CHA₂DS₂-VASc score (the earlier version of the current CHA₂DS₂-VA model) of 0–1 for males and 1–2 for females. Results showed that abelacimab consistently and sustainably reduced FXI levels without causing severe adverse events, including major or clinically relevant bleeding. These findings provided strong support for the continued clinical development of the drug.

The AZALEA-TIMI 71 [Safety and Tolerability of Abelacimab (MAA868) vs. Rivaroxaban in Patients With Atrial Fibrillation] study represents the largest phase 2 trial conducted on abelacimab, aiming to evaluate its safety profile.¹⁸ This was a randomized, open-label study involving 1287 patients with a moderate-to-high risk of cardioembolic ischaemic stroke. Patients were randomized in a 1:1:1 ratio to receive a monthly subcutaneous dose of abelacimab (90 mg or 150 mg) or rivaroxaban (20 mg or 15 mg once a day for patients with creatinine clearance < 50 mL/min, calculated using the Cockcroft–Gault equation). The study’s primary endpoint was the incidence of major bleeding and clinically relevant non-major bleeding according to the criteria of the International Society on Thrombosis and Haemostasis (ISTH). At a median follow-up of 21 months, the incidence of the primary endpoint was 1.9% in the abelacimab 90 mg group, 2.7% in the abelacimab 150 mg group, and 8.1% in the rivaroxaban group (P < 0.001 for both abelacimab doses compared to rivaroxaban). Despite the dosage difference, the degree of FXI inhibition was similar in the two abelacimab groups, reaching 97% with the 90 mg dose and 99% with the 150 mg dose. Due to the clearly favourable safety profile demonstrated by abelacimab, the study was prematurely stopped in September 2023 upon recommendation from the Independent Data Monitoring Committee. However, it is important to note that the number of patients achieving the primary endpoint was lower than planned, reaching only 52% of the expected value outlined in the initial study design. Finally, although a numerically higher (but not statistically significant) rate of ischaemic stroke was observed in patients treated with abelacimab, the rate of all-cause mortality was numerically (though not significantly) lower in the abelacimab groups compared to the rivaroxaban group. Overall, this suggests a favourable risk-benefit profile for the monoclonal antibody.

The phase 2 trial PACIFIC-AF (Proper Dosing and Safety of the Oral FXIa Inhibitor BAY 2433334 in Patients with Atrial Fibrillation) was a randomized, multicentre, double-blind study aimed at determining the optimal dose and evaluating the safety of asundexian in patients with AF at high ischaemic risk and increased bleeding risk.¹⁹ The study enrolled 753 patients aged over 45 years, who were treated for 12 weeks with oral asundexian (20 mg or 50 mg once a day) or apixaban (5 mg twice daily). The primary endpoint was a composite of major bleeding and clinically relevant non-major bleeding, assessed according to ISTH criteria. The results showed a significant reduction in FXI levels after 4 weeks of treatment: −81% at the nadir and −90% at the peak with asundexian 20 mg, and −92% at the nadir and −94% at the peak with asundexian 50 mg. The incidence rates for the primary endpoint compared to apixaban were 0.50 for asundexian 20 mg, 0.16 for the 50 mg dose, and 0.33 for the combined doses. The adverse event rate was similar across the treatment groups (47% with asundexian 20 mg, 47% with 50 mg, and 49% with apixaban), and no major bleeding events were reported. It is important to note that the small sample size limits the ability to draw definitive conclusions regarding the safety and clinical efficacy of asundexian.

The only completed phase 3 study currently on FXI inhibitors in AF is OCEANIC-AF (Oral faCtor Eleven A iNhibitor asundexIan as novel antithrombotic—Atrial Fibrillation study), part of the OCEANIC-Program, conducted on asundexian. The trial was prematurely terminated in November 2023 based on the recommendation of the Independent Data Monitoring Committee due to unfavourable results observed during monitoring. This was a randomized, multicentre, double-blind study aimed at assessing the efficacy and safety of asundexian compared to apixaban in preventing stroke and SE.²⁰ The trial enrolled a total of 14 810 patients with high-risk AF, randomizing them in a 1:1 ratio to receive oral asundexian (50 mg once a day) or apixaban (5 mg or 2.5 mg twice daily, depending on dose-reduction criteria). The primary efficacy endpoint was the incidence of ischaemic stroke and SE, with the primary hypothesis being that asundexian was at least non-inferior to apixaban. The primary safety endpoint was the incidence of major bleeding. The results showed the failure of the primary hypothesis, with an incidence of the primary endpoint of 1.3% in the asundexian group compared to 0.4% in the apixaban group [hazard ratio (HR): 3.79; 95% confidence interval (CI): 2.46–5.83]. However, safety data were favourable for asundexian, with a major bleeding incidence of 0.2% compared to 0.7% for apixaban (HR: 0.32; 95% CI: 0.18–0.55), and with no significant differences in overall adverse events between the groups.

The main characteristics of the studies described in this section are shown in Table 2.

Ongoing studies

Abelacimab is currently being evaluated in the phase 3 LILAC-TIMI 76 trial (Study to Evaluate the Efficacy and Safety of Abelacimab in High-risk Patients With Atrial Fibrillation Who Have Been Deemed Unsuitable for Oral Anticoagulation; NCT05712200). This is a randomized, multicentre, double-blind study that plans to enrol 1900 patients. Inclusion criteria include patients aged 65–74 years with a CHA₂DS₂-VASc score ≥ 4 or those aged ≥75 years with a CHA₂DS₂-VASc score ≥ 3, who are considered unsuitable for OAC therapy and left atrial appendage occlusion, either due to clinical decision or patient refusal. Participants will be randomized in a 1:1 ratio to receive either 150 mg of abelacimab administered subcutaneously once monthly or a placebo. The primary efficacy endpoint is time to the first event of ischaemic stroke or SE, while the primary safety endpoint is time to the first occurrence of type 3c/5 bleeding, according to the Bleeding Academic Research Consortium classification. Both endpoints will be evaluated up to 30 months. The study aims to determine the efficacy and safety of abelacimab in preventing thrombo-embolic events in a high-risk population that has historically been challenging to treat with currently available therapeutic options.

Another phase 3, randomized, multicentre, double-blind, placebo-controlled study, OCEANIC-AFINA (2023-505421-13), has been designed to evaluate the efficacy and safety of oral asundexian in the prevention of stroke in patients aged ≥65 years with AF at high risk of stroke or SE, who are considered ineligible for OAC treatment. The primary efficacy and safety endpoints will be, respectively, time to the first event of ischaemic stroke or SE and time to the first occurrence of major bleeding.

Finally, the LIBREXIA-AF trial (A Study of Milvexian Versus Apixaban in Participants with Atrial Fibrillation; NCT05757869), a randomized, multicentre, double-blind, event-driven study, will evaluate the efficacy and safety of oral milvexian (100 mg twice daily) compared to apixaban (5 mg or 2.5 mg twice daily) in ∼15 500 adult patients with AF. The primary endpoint of the study is time to the first occurrence of the composite of stroke and non-central nervous system SE. Secondary endpoints include time to the first occurrence of major bleeding (evaluated according to ISTH criteria), and time to the first occurrence of the composite of major bleeding and clinically relevant non-major bleeding. The study design estimates 430 events related to the efficacy endpoint and 530 events related to the safety endpoint, with an estimated completion in ∼4 years.

The main characteristics of the studies described in this section are shown in Table 3.

Critical evaluation and future perspectives

The theoretical possibility of achieving a dissociation between the processes of haemostasis and thrombosis through targeted and safe inhibition of FXI, leading to a reduction in the risk of major bleeding, has generated significant expectations for this new class of anticoagulants. The results of phase 2 trials, conducted across different clinical settings, have further fuelled widespread optimism within the scientific community.

In the context of AF, a notable example is the early termination of the AZALEA-TIMI 71 trial, driven by the clear superiority of abelacimab over rivaroxaban in terms of safety. It is important to emphasize, however, that although not statistically significant, a higher number of ischaemic events were observed in the intervention group compared to the control group. While this raises some questions, it does not currently undermine confidence in the potential efficacy of the drug. More conclusive results will emerge from the phase 3 LILAC-TIMI 76 trial, in which abelacimab will be compared to a placebo rather than the DOACs. Although this comparison may seem limiting, the specific clinical context would make the efficacy of abelacimab particularly significant, as it could provide a concrete therapeutic option for elderly AF patients who currently lack an effective treatment for stroke prevention due to being ineligible for OAC therapy.

More advanced data are available for small synthetic molecules, particularly regarding asundexian. Although this drug showed promising characteristics in its phase 2 trial, the premature and unexpected termination of the OCEANIC-AF phase 3 study has raised some degree of skepticism regarding the actual efficacy of FXI inhibitors. The main reason for the termination, as previously mentioned, was the failure of asundexian to demonstrate superiority over apixaban, a result that inevitably raised a series of questions. However, it is essential to emphasize that negative results obtained in a specific clinical context are not automatically generalizable to other conditions. Therefore, the failure of asundexian in AF does not necessarily imply a global failure of FXI inhibitors as a class. Similarly, the negative result of the OCEANIC-AF trial may not reflect the true ineffectiveness of the drug in preventing stroke in AF patients, but rather highlight issues related to the design or planning of the study.

For example, it is worth questioning whether the 50 mg once-daily dose was adequate to provide optimal pharmacological coverage over 24 h for the prevention of thrombotic events. Additionally, it remains unclear whether there is a direct correlation between the degree of FXI inhibition and the observed clinical effects. Another relevant factor is that asundexian was not tested in VTE, a condition that shares several pathophysiological mechanisms with thrombogenesis in AF. This testing could have provided more precise guidance for optimizing the drug dosage.²¹

On the other hand, it is necessary to consider the hypothesis that this pharmacological class might be intrinsically ineffective in the context of AF, as FXI may not represent a relevant biological target in this specific condition. Nevertheless, the termination of the OCEANIC-AF trial should not be interpreted as a definitive obstacle to the development of FXI inhibitors. On the contrary, it should represent an opportunity to gain additional knowledge and improve the design of future clinical studies, starting with the OCEANIC-AFINA trial. The ongoing studies in other clinical settings (e.g. stroke) continue without interruption, even adding new research centres, which strengthens a degree of optimism and confirms scientific interest and confidence in the potential of this pharmacological class.

Among the ongoing studies, the LIBREXIA-AF trial, focused on milvexian, is of particular importance. This study could provide pivotal data on the efficacy of FXI inhibitors in AF. Unlike the OCEANIC-AF study, milvexian in the LIBREXIA-AF trial will be administered in a twice-daily dose, which may ensure more stable drug levels throughout the day, compared to the once-daily administration of asundexian. Furthermore, considering that milvexian is 10 times more potent on a molar basis than asundexian and is being tested at a dose 4 times higher, the results may clarify whether the failure of asundexian was due to an inadequate dosage.²¹ The continuation of the LIBREXIA-AF trial, on the one hand, provides reassurance; however, it is important to maintain cautious optimism, as unexpected developments—both positive and negative—cannot be ruled out.

Ultimately, the currently available data are insufficient to conclusively determine the failure of asundexian and should not represent an insurmountable obstacle to further development of this pharmacological class. A definitive assessment of the efficacy of FXI inhibitors requires waiting for the results of the ongoing trials. Only the publication of these data will determine whether this new therapeutic class represents a false start or a genuine advancement in the treatment of AF.

Conclusions

FXI inhibitors represent a promising innovation in anticoagulant therapy, with the potential to reduce bleeding risk without compromising the prevention of thrombotic events. However, the results regarding the prevention of cardioembolic stroke in patients with AF are conflicting. While reassuring safety data have emerged, the termination of the OCEANIC-AF study on asundexian raises concerns about the efficacy of these drugs in this clinical context. Ongoing studies, LILAC-TIMI 76 and LIBREXIA-AF, will provide decisive data to determine whether these compounds represent a true therapeutic breakthrough or a missed opportunity.

Funding

No funding provided.

Data availability

No new data were generated or analysed in support of this research.

Disclaimer

This paper was originally published in the Italian language as ‘Inibitori del fattore XI e fibrillazione atriale: svolta imminente o falsa partenza?’, in the Volume degli Atti del Congresso “Conoscere e Cuare il Cuore 2025”, published by Centro per la Lotta contro l'Infarto for distribution at the CCC Conference. This paper was translated by Dr. Mario Albertucci, representative of the CLI Foundation, and republished with permission.

References

Author notes