Amrecibart and cenvacibart for the prevention of catheter-related thrombosis (ROXI-CATH): a double-blind, randomised, placebo-controlled, phase 2 trial

Why clinicians should know about it

• Picked for Hematology (paper of the day, 9 November 2026): Randomized trial of catheter thrombosis prevention and bleeding

Abstract

BACKGROUND: Peripherally inserted central catheters (PICCs) are often complicated by venous thromboembolism, yet no thromboprophylaxis is approved. Factor XI (FXI) inhibition might prevent catheter-associated venous thromboembolism without increasing bleeding. We evaluated the safety and efficacy of amrecibart and cenvacibart for preventing venous thromboembolism after PICC placement. METHODS: ROXI-CATH is a double-blind, randomised, placebo-controlled, phase 2 trial conducted across 12 hospital centres in Bulgaria, Canada, Israel, Romania, the UK, and the USA. Eligible participants aged 18 years or older undergoing PICC insertion, with the PICC expected to remain for at least 14 days, and an Eastern Cooperative Oncology Group performance status of 2 or less were randomly assigned (1:1:1) using a permuted block design and stratified by site and diagnosis of active cancer to receive a single intravenous infusion of amrecibart (300 mg; anti-FXI A2 domain monoclonal antibody), cenvacibart (250 mg; anti-FXI catalytic domain monoclonal antibody), or placebo. Participants, principal investigators, and study site staff were masked to treatment assignment. The primary efficacy outcome was adjudicated venous thromboembolism (symptomatic or asymptomatic PICC-associated deep vein thrombosis or superficial vein thrombosis, or pulmonary embolism, including pulmonary embolism-related death) up to day 14. Ultrasounds were performed on days 7 and 14. The key secondary efficacy outcome was adjudicated thrombosis detected by ultrasound in the upper extremity with the PICC up to day 14. The primary safety outcome was site-reported treatment-emergent adverse events up to the end of the study. The key secondary safety outcomes were adjudicated major and clinically relevant non-major (CRNM) bleeding up to day 14. Efficacy analyses were performed in the full analysis set, which consisted of all randomly assigned participants who received any study drug and with any evaluable post-baseline ultrasound or a confirmed episode of venous thromboembolism. Safety analyses were performed in the safety analysis set, which included all randomly assigned participants who received any study drug (as treated). This trial is registered with ClinicalTrials.gov, NCT06299111 (completed; these analyses were performed after database lock on May 6, 2026). FINDINGS: Between Sept 20, 2024, and Dec 19, 2025, 240 participants were screened, and 195 were randomly assigned to amrecibart (n=65), cenvacibart (n=66), or placebo (n=64). One participant in the placebo group who did not receive study treatment and had no evaluable ultrasound was excluded from the full analysis set. The median follow-up time was 100 days (IQR 100-103). The median age was 60 years (49-69), and 89 (46%) of 194 participants were female, 105 (54%) were male, and 192 (99%) were White. Adjudicated venous thromboembolism up to day 14 occured in 13 (20%) of 65 participants in the amrecibart group (odds ratio [OR] 0·37 [90% credible interval 0·18-0·65]), 13 (20%) of 66 participants in the cenvacibart group (0·36 [0·17-0·64]) compared with 27 (43%) of 63 participants in the placebo group. Adjudicated PICC-associated venous thromboembolism up to day 14 occurred in 12 (18%) participants in the amrecibart group (OR 0·30 [95% CI 0·14-0·67]), 13 (20%) participants in the cenvacibart group (0·32 [0·14-0·69]), compared with 27 (43%) participants in the placebo group. Site-reported treatment-emergent adverse events up to day 100 occurred in 45 (69%) participants in the amrecibart group, 42 (64%) participants in the cenvacibart group, and 48 (76%) participants in the placebo group. No participants had major bleeding up to day 14. CRNM bleeding up to day 14 occurred in two (3%) participants in the amrecibart group, two (3%) in the cenvacibart group, and two (3%) in the placebo group. Serious adverse events occurred in nine (14%) participants in the ambrecibart group, six (9%) in the cenvacibart group, and ten (16%) in the placebo group. One (2%) participant in the amrecibart group, four (6%) participants in the cenvacibart group, and two (3%) participants in the placebo group died; these deaths were assessed as not related to treatment by the respective investigators. INTERPRETATION: A single infusion of amrecibart or cenvacibart reduced the incidence of venous thromboembolism following PICC insertion, suggesting FXI inhibition at the A2 or catalytic domain as a promising prophylactic strategy for contact-mediated thrombosis. FUNDING: Regeneron Pharmaceuticals.

Abstract as published, via PubMed.