1. Publication
- Title: Efficacy and safety of REGN9933A2 and REGN7508Cat for preventing postoperative venous thromboembolism (ROXI-VTE-I and ROXI-VTE-II): two randomised, open-label, phase 2 trials
- Acronym: ROXI-VTE-I/II
- Year & Journal: The Lancet, epublished November 8, 2025 (print: 2025;406(10519):2551-2563)
- Citation: Weitz JI, Kithcart AP, O'Brien MP, et al. Lancet. 2025;406(10519):2551-2563. doi:10.1016/S0140-6736(25)02097-5
2. Context & Rationale
Background: Current anticoagulants for postoperative VTE prevention carry bleeding risk that limits their use. Factor XI (FXI) inhibitors may reduce thrombosis without increasing bleeding, since FXI deficiency is associated with lower thrombotic risk without excess bleeding tendency.
Research Question/Hypothesis: In patients undergoing knee arthroplasty, are two mechanistically distinct FXI-targeting antibodies (REGN9933A2, apple-2-domain; REGN7508Cat, catalytic-domain) superior to enoxaparin for VTE prevention?
3. Design & Methods
- Study Type: Two randomized, open-label, active-controlled phase 2 trials
- Setting & Centers: ROXI-VTE-I: 15 centers/7 countries; ROXI-VTE-II: 12 centers/5 countries
- Population: Adults ≥50y undergoing unilateral total knee arthroplasty
- Intervention: ROXI-VTE-I: single IV REGN9933A2 300mg (n=116) vs enoxaparin (n=117) vs apixaban (n=113, exploratory). ROXI-VTE-II: single IV REGN7508Cat 250mg vs enoxaparin
- Randomization: ROXI-VTE-I n=373 (May 2023–May 2024); ROXI-VTE-II n=179 (Jun 2024–Jan 2025)
- Statistical Power & Follow-Up: Bayesian primary analysis (superiority if posterior probability log OR<0 >95%). Primary: venogram-detected asymptomatic or confirmed symptomatic VTE. Safety: major + clinically relevant non-major bleeding.
4. Key Results
Outcome | REGN9933A2 | Enoxaparin | Apixaban | Effect Size | Notes |
VTE rate, ROXI-VTE-I (primary) | 17.2% (20/116) | 22.2% (26/117) | 12.4% (14/113) | Not superior to enoxaparin | REGN9933A2 did not meet superiority threshold |
VTE rate, ROXI-VTE-II (REGN7508Cat vs enoxaparin) | — | — | — | Adjusted OR 0.37, 95% CI 0.2–0.68 | REGN7508Cat superior to enoxaparin |
Bottom line: Only REGN7508Cat (catalytic-domain antibody) demonstrated superiority to enoxaparin; REGN9933A2 (apple-2-domain antibody) did not.
5. Internal Validity Assessment
Rigorous Bayesian phase 2 design with prespecified superiority thresholds; open-label but objective venographic primary outcome. Overall: Moderate-to-strong — credible phase 2 proof-of-concept data, though these are still early-phase, industry-sponsored (Regeneron) trials requiring phase 3 confirmation.
6. External Validity Assessment
International, multi-country knee arthroplasty population — a standard VTE-prevention model population, though results in this elective-surgery context may not generalize to other VTE-prevention indications.
7. Strengths & Limitations
Strengths: Tests two mechanistically distinct FXI antibodies, informative for future drug development; Bayesian design efficient for early-phase dose-finding/proof-of-concept.
Limitations: Open-label; industry-sponsored; phase 2 only — not yet confirmatory; REGN9933A2 failed its primary superiority hypothesis.
8. Interpretation & Practice Impact
Supports continued development of REGN7508Cat (but not REGN9933A2) as a novel FXI-targeting anticoagulant for VTE prevention — an early but genuine step toward potentially safer (lower-bleeding-risk) anticoagulation, pending phase 3 confirmation.
9. Controversies & Subsequent Evidence
Accompanying Lancet commentary ("New light on the activators of factor XI in venous thromboembolism") frames this as advancing understanding of FXI activation pathways (FXIIa vs thrombin-mediated) and their differential clinical relevance — the two-antibody, two-trial design usefully illustrates that not all FXI-targeting approaches are equally effective.
10. Summary & Executive Takeaway
Summary: Two phase 2 trials (n=373 and n=179) tested two distinct FXI antibodies against enoxaparin for VTE prevention after knee arthroplasty. REGN7508Cat (catalytic-domain) was superior to enoxaparin (adjusted OR 0.37, 95% CI 0.2-0.68), while REGN9933A2 (apple-2-domain) was not superior (17.2% vs 22.2% VTE).
Overall Takeaway: Factor XI inhibition remains a promising anticoagulation strategy, but this early-phase data shows mechanism matters — catalytic-domain blockade (REGN7508Cat) demonstrated clear efficacy while apple-2-domain blockade (REGN9933A2) did not, providing important mechanistic guidance for the next generation of FXI-targeted anticoagulant development.
11. Bibliography
- Hunt BJ, et al. New light on the activators of factor XI in venous thromboembolism [commentary]. Lancet. 2025.