Factor XI Inhibitors or Direct Oral Anticoagulants in Atrial Fibrillation: A Focused Meta-analysis

Direct oral anticoagulants (DOACs) are the preferred anticoagulants across various clinical settings, including atrial fibrillation (AF). However, their use is inevitably associated with an increased risk of bleeding, which carries important prognostic implications.[1] Targeting coagulation factors involved in the intrinsic pathway, such as factor XI (FXI), has been proposed to reduce bleeding while maintaining adequate ischemic protection.[2] Indeed, the intrinsic pathway plays a crucial role in the amplification phase of coagulation, driving the formation of large thrombi and pathological thrombosis, while playing a lesser role in physiological hemostasis.[2] Early phase II trials comparing FXI inhibitors versus DOACs have shown encouraging results, suggesting an improved safety profile of FXI inhibitors. However, these studies were inconclusive regarding their antithrombotic efficacy, due to the low statistical power and the use of mixed doses of FXI inhibitors.[3] A meta-analysis of early phase II trials, stratified by FXI inhibitor dose, raised concerns about whether FXI inhibitors achieve noninferiority in efficacy compared to DOACs across different clinical settings.[4]

In the setting of AF, two randomized trials have recently become available, the OCEANIC-AF and the AZALEA-TIMI 71.[5] [6] The OCEANIC-AF was a phase III, randomized, double-blind, noninferiority trial randomizing 14,830 patients to receive asundexian 50 mg once daily or apixaban 5 mg twice daily. The primary efficacy endpoint was stroke or systemic embolism, and the primary safety endpoint was International Society on Thrombosis and Hemostasis (ISTH) major bleeding events. Although there was a reduction in bleeding, the trial was stopped prematurely due to a numerically higher incidence of stroke or systemic embolism with asundexian.[6] Mean follow-up was 155 days. AZALEA-TIMI 71 was a phase II, open label trial randomizing 1,287 patients to receive subcutaneous injection of abelacimab 150 or 90 mg once monthly administered in a blinded fashion or oral rivaroxaban 20 mg once daily. The primary endpoint was major or clinically relevant ISTH non-major bleeding at 657 days.[5] Abelacimab reduced bleeding events but was associated with a nonsignificant numerical increase in stroke compared with rivaroxaban.[5] These two trials complement previous evidence from the phase II, double-blind PACIFIC-AF trial, which randomized 862 patients to receive asundexian 20 mg once daily, 50 mg once daily, or apixaban 5 mg twice daily, being followed up for 84 days.[7] However, it is important to underscore that none of these trials were powered for efficacy outcomes. Meta-analyses may overcome this limitation by increasing statistical power. Against this background, We conducted a meta-analysis of randomized trials comparing FXI inhibitors versus DOACs in patients with AF.

The present study was conducted in agreement with PRISMA and Cochrane's collaboration recommendations, and registered in PROSPERO (CRD420251088578). Trials enrolling patients with AF who were randomly assigned to receive FXI inhibitor or DOAC and reporting at least one outcome of interest were included. Screening of three different databases and data extraction were conducted by two experienced reviewers (C.L. and M.G.), with discrepancies resolved by consensus. The primary efficacy endpoint was stroke, and the primary safety endpoint was ISTH-defined major bleeding. Any bleeding, trial-defined major adverse cardiac event (MACE), and intracranial hemorrhages were the key secondary endpoints. To adjust for varying follow-up durations across trials, trial follow-up-weighted incidence rate ratios (IRRs) with 95% confidence intervals (CIs) were pooled by recalculating patient-years based on each trial's maximum available follow-up using study-level contrast frequentist fixed- and random-effect models, depending on heterogeneity. Additionally, analyses were conducted within pre-specified subgroups based on the use of high-dose versus low-dose FXI inhibitors. Leave-one-out analysis was performed as sensitivity analysis. Heterogeneity was measured using τ² and I² statistics. Statistical analysis was conducted through R 4.3.1.

Three trials were included (PACIFIC-AF, OCEANIC-AF, and AZALEA-TIMI 71),[5] [6] [7] encompassing a total of 16,845 patients and 9,594 patient-year. Among included trials, mean CHA₂DS₂-VASc score and follow-up durations were 4.4 and 298.7 days, respectively. Compared with DOACs, FXI inhibitors reduced major bleeding (IRR 0.30; 95% CI 0.21–0.45) but were associated with an increase in stroke (IRR 1.36; 95% CI 1.04–1.77; [Fig. 1]). Although there was no significant interaction between treatment effect estimates and FXI inhibitor doses, the increase in stroke was numerically higher with the lower dosages (IRR 2.41, 95% CI 0.85–6.86) compared with the higher dosages (IRR 1.30, 95% CI 0.99–1.71; p_int = 0.25). FXI inhibitors reduced any bleeding (IRR 0.52; 95% CI 0.44–0.61) without significantly affecting MACE (IRR 1.21; 95% CI 0.77–1.91) or intracranial hemorrhages (IRR 0.42; 95% CI 0.13–1.29). No major influence of one trial over another for treatment effect estimates was detected for any of the endpoints.

Fig. 1 Forest plot for primary endpoints with interaction for drug dosage. CI, confidence interval; IRR, incidence rate ratio; P _int, P for interaction.

This is the first meta-analysis comparing FXI inhibitors versus DOACs in AF patients that includes the recent OCEANIC-AF and AZALEA-TIMI 71 trials. This analysis confirms the superior safety profile of FXI inhibitors versus DOACs for any and major bleeding. However, it also raises concerns regarding the antithrombotic efficacy of FXI inhibitors for stroke prevention. Although dose-dependent interaction analysis suggests lack of dose influence on overall results, further investigation into optimal dosing strategies is warranted, given the numerically higher rates of stroke with low compared to higher doses of FXI inhibitors and the modest statistical power for subgroup analyses. Indeed, previous studies highlighted dose-dependent differences in the pharmacodynamic and pharmacokinetic profiles of FXI inhibitors, including their inhibition constants and mechanisms of action.[8] Phase 3 trials with adequate statistical power for efficacy, evaluating FXI inhibitors at varying doses or with distinct mechanisms of action are necessary to establish their role in this clinical setting. The results of the LIBREXIA-AF trial (NCT05757869), which is comparing milvexian (100 mg orally, twice daily) to apixaban (5 mg twice daily) in approximately 20,000 patients with AF, are expected by 2027.

Our results should be seen in light of several limitations. First, the lack of individual patient level-data does not allow examination of specific confounders on the pooled estimate, although sensitivity analyses partially mitigated this concern. Second, the IRR relied on declared maximum follow-ups, given the lack of patient-level follow-up data, which did not account for attrition of early events across trials. Lastly, the inclusion of small studies may have introduced some small-study effect, although leave-one-out analysis did not detect any influence of one study over another.

Data Availability Statement

All data are available upon reasonable request to the corresponding author.

Publikationsverlauf

Eingereicht: 17. März 2025

Angenommen: 01. August 2025

Accepted Manuscript online:
04. August 2025

Artikel online veröffentlicht:
12. August 2025

• References

• 1 Galli M, Gragnano F, Berteotti M. et al; Working Group of Thrombosis of the Italian Society of Cardiology. Antithrombotic therapy in high bleeding risk, part I: percutaneous cardiac interventions. JACC Cardiovasc Interv 2024; 17 (19) 2197-2215
• 2 Occhipinti G, Laudani C, Spagnolo M. et al. Pharmacological and clinical appraisal of factor XI inhibitor drugs. Eur Heart J Cardiovasc Pharmacother 2024; 10 (03) 245-258
• 3 Galli M, Occhipinti G, Ortega-Paz L. et al. Therapeutic potential of FXI inhibitors: hype or hope?. Drugs 2024; 84 (09) 1055-1070
• 4 Galli M, Laborante R, Ortega-Paz L. et al. Factor XI inhibitors in early clinical trials: a meta-analysis. Thromb Haemost 2023; 123 (06) 576-584
• 5 Ruff CT, Patel SM, Giugliano RP. et al; AZALEA–TIMI 71 Investigators. Abelacimab versus rivaroxaban in patients with atrial fibrillation. N Engl J Med 2025; 392 (04) 361-371
• 6 Piccini JP, Patel MR, Steffel J. et al; OCEANIC-AF Steering Committee and Investigators. Asundexian versus apixaban in patients with atrial fibrillation. N Engl J Med 2025; 392 (01) 23-32
• 7 Piccini JP, Caso V, Connolly SJ. et al; PACIFIC-AF Investigators. Safety of the oral factor XIa inhibitor asundexian compared with apixaban in patients with atrial fibrillation (PACIFIC-AF): a multicentre, randomised, double-blind, double-dummy, dose-finding phase 2 study. Lancet 2022; 399 (10333): 1383-1390
• 8 Spagnolo M, Capodanno D. Factor XI Inhibitors for Myocardial Infarction: A Safer Path to Thrombotic Protection. Thromb Haemost 2025; (e-pub ahead of print).