Andexanet alfa vs non-specific treatments for intracerebral hemorrhage in patients taking factor xa inhibitors - individual patient data analysis of annexa-4 and tich-noac

BACKGROUND:Data comparing the specific reversal agent andexanet alfa with non-specific treatments in patients with non-traumatic intracerebral hemorrhage (ICH) associated with factor-Xa inhibitor (FXaI) use are scarce. AIM:To determine the association between use of andexanet alfa (compared to non-specific treatments) with rate of hematoma expansion and thromboembolic complications in patients with FXaI-associated ICH. METHODS:We performed an individual patient data analysis combining two independent, prospective studies: ANNEXA-4 (180 patients receiving andexanet alfa, NCT02329327) and TICH-NOAC (63 patients receiving tranexamic acid or placebo +/- prothrombin complex concentrate, NCT02866838). The primary efficacy outcome was hematoma expansion on follow-up imaging. The primary safety outcome was any thromboembolic complication (ischemic stroke, myocardial infarction, pulmonary embolism or deep vein thrombosis) at 30 days. We used binary logistic regression models adjusted for baseline hematoma volume, age, calibrated anti-Xa activity, times from last intake of FXaI and symptom onset to treatment, respectively. RESULTS:Among 243 participants included, the median age was 80 years (IQR 75-84), baseline hematoma volume was 9.1ml (IQR 3.4-21) and anti-Xa activity 118ng/ml (IQR 78-222). Times from last FXaI intake and symptom onset to treatment were 11 hours (IQR 7-16) and 4.7 hours (IQR 3.0-7.6), respectively. Overall, 21% (n=50) of the patients experienced hematoma expansion (ANNEXA-4: 13%, n=24; TICH-NOAC: 41%, n=25). After adjusting for pre-specified confounders (baseline hematoma volume, age, calibrated anti-Xa activity, times from last intake of FXaI and symptom onset to treatment, respectively), treatment with andexanet alfa was independently associated with decreased odds for hematoma expansion (aOR 0.33, 95%CI 0.13-0.80, p=0.015). Overall, 11% (n=26) of patients had any thromboembolic complication within 30 days (ANNEXA-4: 11%, n=20; TICH-NOAC: 10%, n=6). There was no association between any thromboembolic complication and treatment with andexanet alfa (aOR 0.70, 95%CI 0.16-3.12, p=0.641). CONCLUSIONS:Use of andexanet alfa compared to any other non-specific treatment strategy was associated with decreased odds for hematoma expansion, without increased odds for thromboembolic complications.