Abelacimab versus Rivaroxaban in Patients with Atrial Fibrillation (AZALEA–TIMI 71)

Authors, Journal, Affiliations, Type, DOI

• Authors: Christian T. Ruff, Siddharth M. Patel, Robert P. Giugliano, David A. Morrow, Bruce Hug, Julia F. Kuder, Erica L. Goodrich, Shih-Ann Chen, Shaun G. Goodman, Boyoung Joung, Robert G. Kiss, Jindrich Spinar, Wojciech Wojakowski, Jeffrey I. Weitz, Sabina A. Murphy, Stephen D. Wiviott, Sanobar Parkar, Daniel Bloomfield, Marc S. Sabatine, for the AZALEA–TIMI 71 Investigators
• Journal: N Engl J Med 2025;392(4):361–71. Published in the 23 January 2025 issue
• Affiliations: TIMI Study Group, Brigham and Women's Hospital, Harvard Medical School, Boston (lead); Anthos Therapeutics, Cambridge, MA; 95 centres in seven countries, with participant sites concentrated in Europe (70.9%), North America (24.0%) and Asia (5.1%)
• Type: Multinational, phase 2b, parallel-group, partially blinded (dose-blinded within abelacimab arms, open-label abelacimab versus rivaroxaban), randomised, active-controlled safety trial. Funded by Anthos Therapeutics; analyses performed independently by the TIMI Study Group from a raw database supplied by the sponsor. Registered NCT04755283
• DOI: 10.1056/NEJMoa2406674

Overview

AZALEA–TIMI 71 asked whether a factor XI inhibitor could match a direct oral anticoagulant's stroke protection while bleeding less. In 1,287 patients with AF and a median CHA₂DS₂-VASc of 5, monthly subcutaneous abelacimab at 150 mg or 90 mg reduced median free factor XI levels by 99% and 97% respectively and cut the primary end point of major or clinically relevant nonmajor bleeding by 62% and 69% relative to rivaroxaban (3.22 and 2.64 versus 8.38 events per 100 person-years). The trial was stopped early on 14 September 2023 by its independent data monitoring committee, after a median 1.8 years, because the benefit was larger than anticipated. The bleeding result is large and internally consistent — a dose–response relationship, a near-elimination of gastrointestinal bleeding, and a net clinical outcome favouring abelacimab — but the trial was neither designed nor powered to show that stroke prevention is preserved, and ischemic strokes were numerically more frequent on abelacimab. That ambiguity was resolved in the negative nine months later by OCEANIC-AF, which tested a different factor XIa inhibitor against apixaban for efficacy and failed.

Keywords

Not provided — NEJM research articles do not carry author keywords. Indexing terms: atrial fibrillation; factor XI; factor XIa; abelacimab; rivaroxaban; direct oral anticoagulant; bleeding; major bleeding; clinically relevant nonmajor bleeding; gastrointestinal bleeding; net clinical outcome.

Key Takeaways

Background — why factor XI

• AF affects an estimated 50 million people worldwide and carries a five-fold stroke risk; anticoagulation is the cornerstone of prevention, and guidelines prefer DOACs over VKAs.
• Bleeding, especially gastrointestinal bleeding, remains the major complication of DOAC therapy and a leading reason patients go undertreated. A safer anticoagulant has therefore been the goal.
• Factor XI is essential for thrombosis but largely dispensable for haemostasis: it acts in the amplification phase (activated by thrombin after initiation) rather than the initiation phase. People with inherited factor XI deficiency have fewer embolic events without appreciable spontaneous bleeding.
• The therapeutic hypothesis is therefore to uncouple thrombosis from haemostasis rather than to trade one against the other.

Abelacimab — agent and pharmacology

• Fully human monoclonal antibody that binds the catalytic domain of factor XI and locks it in the inactive state.
• Dual mechanism: it blocks activation of factor XI and inhibits factor XIa directly — a distinction from small molecules that inhibit factor XIa activity only.
• Intravenous dosing suppresses factor XI almost completely within 1 hour, with near-maximal inhibition sustained up to 30 days, which is what permits once-monthly subcutaneous dosing.
• Prior proof of concept: ANT-005, in elective knee arthroplasty, showed an 80% reduction in venous thromboembolism versus enoxaparin without excess bleeding with a single 150 mg intravenous dose.

Design and population

• 1,287 patients randomised 1:1:1, March–December 2021, at 95 centres in seven countries.
• Abelacimab 150 mg SC monthly, abelacimab 90 mg SC monthly (double-blind to dose), or rivaroxaban 20 mg daily open-label (15 mg if creatinine clearance ≤50 mL/min).
• Entry: age ≥55, documented AF or flutter, planned anticoagulation, and CHA₂DS₂-VASc ≥4 — or ≥3 with planned antiplatelet therapy or creatinine clearance ≤50 mL/min.
• Median age 74; 44% women; median CHA₂DS₂-VASc 5; median HAS-BLED 2–3; 47.9% heart failure; 97% hypertension; 24.7% planned concomitant antiplatelet therapy.
• 92% had received an anticoagulant for ≥60 days before entry, 66% of those a DOAC — so this was largely an anticoagulation-experienced population, not an OAC-naive one.
• Primary end point: composite of major or clinically relevant nonmajor bleeding by ISTH criteria, in the on-treatment population (within 60 days of last dose).

Pharmacodynamics

• At 3 months, median reduction in free factor XI was 99% (IQR 98–99) with 150 mg and 97% (IQR 51–99) with 90 mg — the wide interquartile range at 90 mg indicates that the lower dose did not reliably achieve near-complete suppression in every patient.

Bleeding outcomes (complete data set)

• Major bleeding occurred in 2.3% (150 mg), 1.9% (90 mg) and 7.2% (rivaroxaban).
• Major gastrointestinal bleeding was almost abolished: two events in each abelacimab arm versus 18 with rivaroxaban. The authors note the gastrointestinal tract is the commonest bleeding site on DOACs and the one type of bleeding where DOACs do not beat warfarin, and suggest local gastrointestinal exposure to an oral direct-acting agent may itself contribute to bleeding risk — a parenteral agent would avoid it.

Stroke, embolism and death

• 28 patients had stroke or systemic embolism. Overall event rates were low (~1% per year), but the excess in both abelacimab arms was driven by ischemic stroke, with hazard ratios above 2 and wide confidence intervals crossing 1.
• Net clinical outcome was a composite of ischemic stroke, systemic embolism, major or CRNM bleeding, or death from any cause.
• The composite of death or any stroke was similar across arms (HR 0.90 for 150 mg; 1.07 for 90 mg).

Safety

• Adverse events were similar across groups: any 83.8% (150 mg), 82.6% (90 mg), 81.3% (rivaroxaban); serious 36.8%, 37.2%, 39.0%; leading to discontinuation 6.8%, 7.5%, 6.8%.
• Injection-site reactions occurred in 2.8% (150 mg) and 1.6% (90 mg).
• No antidrug antibodies developed in any patient.

Interpreting AZALEA against OCEANIC-AF

• The authors pre-emptively distinguish abelacimab from asundexian on three grounds:
  1. Mechanism — abelacimab binds inactive factor XI and prevents generation of factor XIa, whereas asundexian inhibits factor XIa activity only. Whether upstream or combined inhibition is required for stroke prevention is unknown.
  2. Potency — abelacimab achieves 99% reduction in free factor XI at the 150 mg dose used in phase 3, whereas asundexian 50 mg appears less potent.
  3. Proof of concept — abelacimab demonstrated antithrombotic efficacy against enoxaparin in ANT-005; no equivalent proof-of-concept trial was performed with asundexian.
• The paper is explicit that "the results of ongoing phase 3 trials are necessary to establish the benefit-to-risk ratio of abelacimab."
• LILAC–TIMI 76 (NCT05712200) is testing abelacimab against placebo in patients with AF in whom current anticoagulation is deemed unsuitable — a fundamentally different question from AZALEA's, and one that avoids having to match a DOAC.

Limitations of the document

Authors' own stated limitations

• Open-label randomisation to abelacimab versus rivaroxaban; bias was minimised by dose-blinding within the abelacimab arms and by central adjudication by a committee unaware of assignment, but the comparison itself was unblinded.
• Small sample size precludes assessment of clinical efficacy — the trial had 80% power to detect a 40% reduction in bleeding, not to exclude a stroke difference.
• Generalisability — 94–95% of participants were White.
• Comparator restriction — the trial cannot define bleeding risk relative to DOACs other than rivaroxaban.

Additional limitations evident in the design

• Two different effect estimates are reported. The primary analysis in the abstract and Table 2 uses the complete data set (HR 0.38 / 0.31). The narrative Results section also reports the data-cutoff analysis at the time the monitoring committee recommended termination: 2.69 and 1.87 versus 8.14 events per 100 person-years, HR 0.33 (0.19–0.55) and 0.23 (0.13–0.42). The effect was therefore larger at the stopping analysis and attenuated with fuller follow-up. Which numbers get quoted matters: the more favourable HRs are the interim ones.
• The rivaroxaban bleeding rate was high. Major or CRNM bleeding of 8.38/100 person-years and major bleeding of 3.73/100 person-years exceed most contemporary DOAC trials, in a population with median HAS-BLED 2–3 and 24.7% on concomitant antiplatelet therapy. Part of the relative benefit may be a property of the comparator arm rather than of abelacimab.
• The primary end point is a safety end point, so a drug that prevents no strokes can win it. This is the structural weakness the trial itself acknowledges and OCEANIC-AF later exploited.
• Early termination inflates apparent benefit and truncates follow-up; the median exposure was 2.1 years, which is short for a drug intended for indefinite use.
• Dose selection is unresolved. The 90 mg arm bled less than 150 mg but achieved a median of only 97% factor XI suppression with a lower quartile of 51%, and the 150 mg dose is the one taken forward — so the dose with the best bleeding profile is not the dose being developed.
• No active stroke-prevention comparator with adequate power means the ischemic stroke signal cannot be distinguished from chance, but neither can it be dismissed: HR >2 in both arms, with the point estimate driven by ischemic rather than haemorrhagic events.

Key Concepts Mentioned

• concepts/Factor-XI-Inhibition — the class this trial defines; AZALEA is the safety anchor and the source of the bleeding estimates
• concepts/NOAC-AF-Management — the DOAC standard against which the class is measured
• entities/Atrial-Fibrillation — the indication
• Factor XI / factor XIa; contact activation pathway; amplification phase of coagulation
• Uncoupling thrombosis from haemostasis — the pharmacological hypothesis
• ISTH major and clinically relevant nonmajor bleeding definitions
• Net clinical outcome composites
• Phase 2b dose-finding versus phase 3 confirmatory design

Key Entities Mentioned

• entities/Atrial-Fibrillation
• Abelacimab (MAA868) — Anthos Therapeutics
• Rivaroxaban — the active comparator
• Asundexian — the factor XIa inhibitor that failed OCEANIC-AF, discussed throughout the Discussion
• TIMI Study Group — independent analysis
• Anthos Therapeutics — sponsor
• LILAC–TIMI 76 — ongoing phase 3, abelacimab versus placebo

Wiki Pages Updated

• Created: wiki/sources/abelacimab-azaleatimi71-nejm-2025.md
• Created: wiki/concepts/Factor-XI-Inhibition.md
• Updated: wiki/concepts/NOAC-AF-Management.md
• Updated: wiki/entities/Atrial-Fibrillation.md
• Updated: wiki/sourceindex.md
• Updated: wiki/wikiindex.md