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

BACKGROUND: Coagulation factor XI (FXI) inhibitors can reduce the incidence of thrombosis without increasing bleeding risk. FXI is activated by factor XIIa (FXIIa) or thrombin. REGN7508^Cat is an antibody that binds to the FXI catalytic domain, blocking both its activity and activation by FXIIa and thrombin. REGN9933^A2 binds to the FXI apple 2 domain and blocks FXI activation by FXIIa. We aimed to compare the efficacy and safety of REGN9933^A2 and enoxaparin, with apixaban as an exploratory comparator, and REGN7508^Cat and enoxaparin for venous thromboembolism prevention. METHODS: ROXI-VTE-I and ROXI-VTE-II are randomised, open-label, phase 2 studies in patients (aged ≥50 years) undergoing knee arthroplasty across 15 centres in seven countries and 12 centres in five countries, respectively. In ROXI-VTE-I, patients were randomised (1:1:1) to receive REGN9933^A2 (300 mg intravenously once), enoxaparin (40 mg subcutaneously once daily), or apixaban (2·5 mg orally twice a day; an exploratory comparator) after surgery. In ROXI-VTE-II, patients were randomised (2:1) to receive REGN7508^Cat (250 mg intravenously once) or enoxaparin (40 mg subcutaneously once daily) after surgery. Enoxaparin and apixaban were continued through the day of venography or day 12, whichever was earlier. The primary endpoint in both studies was objectively confirmed venous thromboembolism (a composite of asymptomatic deep-vein thrombosis of the operated leg, objectively confirmed symptomatic deep-vein thrombosis of either leg, or confirmed non-fatal or fatal pulmonary embolism) through day 12 after the first dose of study drug (administered 12-24 h after the end of surgery). In the modified intention-to-treat population (randomly allocated patients who received at least one dose of study treatment and had either an evaluable venogram or confirmed symptomatic venous thromboembolism) for the primary analyses, REGN9933^A2 or REGN7508^Cat was considered superior to enoxaparin if the posterior probability that the log odds ratio (OR) <0 was greater than 95%. For ROXI-VTE-II, the OR was estimated for REGN7508^Cat versus the combined enoxaparin groups of ROXI-VTE-I (discounted by half) and ROXI-VTE-II. The main safety outcome was the composite of major and clinically relevant non-major bleeding. ROXI-VTE-I and ROXI-VTE-II are registered with ClinicalTrials.gov (NCT05618808 and NCT06454630). FINDINGS: Patients were enrolled from May 24, 2023, to May 27, 2024, in ROXI-VTE-I (n=373) and from June 27, 2024, to Jan 21, 2025, in ROXI-VTE-II (n=179). The median follow-up time was 74 days (IQR 72-76) for both studies. In ROXI-VTE-I, venous thromboembolism occurred in 20 (17%) of 116 patients in the REGN9933^A2 group, 26 (22%) of 117 in the enoxaparin group, and 14 (12%) of 113 in the apixaban group through day 12. For REGN9933^A2 versus enoxaparin, the mean adjusted OR (aOR) was 0·78 (90% credible interval 0·47-1·32); the posterior probability was 78·5%. In ROXI-VTE-II, venous thromboembolism occurred in eight (7%) of 113 in the REGN7508^Cat group and ten (17%) of 58 patients in the enoxaparin group through day 12. For REGN7508^Cat versus the combined enoxaparin groups, the mean aOR was 0·37 (0·20-0·68); the posterior probability was 99·8%. There were no major or clinically relevant non-major bleeds in any group. The most common treatment-emergent adverse event was postoperative anaemia (nine [7%] of 123 patients in the REGN9933^A2 group, 11 [9%] of 125 in the enoxaparin group, and 16 [13%] of 125 in the apixaban group in ROXI-VTE-I; and six [5%] of 120 in the REGN7508^Cat group in ROXI-VTE-II). Serious adverse events occurred in four (3%), one (1%), and two (2%) patients given REGN9933^A2, enoxaparin, and apixaban in ROXI-VTE-I; and in two (2%) patients who received REGN7508^Cat and none who received enoxaparin in ROXI-VTE-II. No treatment-related deaths occurred. INTERPRETATION: REGN7508^Cat was superior to enoxaparin for venous thromboembolism prevention; REGN9933^A2 was not superior to enoxaparin. The findings with REGN7508^Cat confirm the role of FXI in postoperative venous thromboembolism, and the findings with REGN9933^A2 suggest that FXIIa-driven FXI activation contributes to this process. FUNDING: Regeneron Pharmaceuticals.