PDF: ANNEXA-I
(2024)Objective
To assess the efficacy and safety of andexanet alfa in reversing anticoagulation in patients with factor Xa inhibitor-associated intracerebral hemorrhage (ICH).
Study Summary
• Hemostatic efficacy (good or excellent): 82% andexanet vs 79% usual care (difference 3.7%, 95% CI −3.3 to 10.7; p=0.29, non-significant)
• Excellent functional outcome (mRS 0–3) at 30 days: 55% vs 53%; 30-day thrombotic events: 10% vs 6%; no difference in mortality (15% vs 16%)
Intervention
Open-label trial comparing standard care vs. andexanet alfa for reversal of apixaban or rivaroxaban in acute ICH.
Inclusion Criteria
Patients with acute ICH within 6 hours of onset who were on apixaban or rivaroxaban, with elevated anti-factor Xa levels.
Study Design
Arms: Andexanet Alfa vs. Usual Care
Patients per Arm: Andexanet: 254, Usual Care: 254
Outcome
• Hemostatic efficacy: 82% (andexanet) vs. 79% (usual care).
• Excellent mRS 0–1 at 30 days: 16% vs. 14%.
• Mortality: 27% in both arms.
Bottom Line
Andexanet alfa, compared to usual care, significantly reduced the median percentage increase in hematoma volume and was associated with a higher rate of excellent or good hemostatic efficacy in patients with acute intracerebral hemorrhage receiving factor Xa inhibitors. However, it led to a higher rate of thrombotic events and death.
Major Points
- 530 patients with acute intracerebral hemorrhage and factor Xa inhibitor use were randomized (265 to andexanet, 265 to usual care).
- Enrollment was stopped early after the first interim analysis due to clear evidence of the superiority of andexanet alfa for the primary outcome.
- The median percentage increase in hematoma volume from baseline to 12 hours was 6.7% in the andexanet group vs 25.1% in the usual-care group (adjusted ratio of geometric means, 0.27; 95% Bayesian credible interval, 0.17 to 0.43; posterior probability of superiority, >0.999).
- Excellent or good hemostatic efficacy was achieved in 80% in the andexanet group vs 62% in the usual-care group (adjusted risk difference, 18 percentage points; 95% Bayesian credible interval, 11 to 26; posterior probability of superiority, >0.999).
- The rate of thrombotic events was 10.3% in the andexanet group vs 2.6% in the usual-care group (adjusted risk ratio, 4.08; 95% CI, 2.05 to 8.12; P<0.001).
- Death from any cause occurred in 24.4% in the andexanet group vs 19.5% in the usual-care group (adjusted risk ratio, 1.25; 95% CI, 0.92 to 1.71; P=0.17).
- Andexanet alfa was initiated a median of 4.4 hours after hemorrhage onset.
Study Design
- Study Type
- Multicenter, randomized, open-label, blinded-endpoint, superiority trial (Phase 3b/4)
- Randomization
- Yes
- Blinding
- Blinded-endpoint design (assessment of primary and secondary outcomes by an independent central adjudication committee blinded to treatment assignment).
- Sample Size
- 530
- Follow-up
- 90 days (primary outcome) and 1 year (secondary outcome).
- Centers
- 168
- Countries
- 22 countries
Primary Outcome
Definition: Median percentage increase in hematoma volume from baseline to 12 hours.
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 25.1% | 6.7% | 0.27 (0.17 to 0.43 (ratio of geometric means)) | >0.999 (posterior probability of superiority) |
Limitations & Criticisms
- The trial was stopped early after the first interim analysis due to overwhelming evidence of superiority for the primary outcome, which means it did not reach its full planned sample size. This limits the power of secondary outcome analyses and safety evaluations.
- The study identified a significantly higher rate of thrombotic events and a numerically higher rate of death in the andexanet alfa group, which are significant safety concerns despite the hematoma volume reduction.
- The trial was open-label, meaning patients and treating clinicians were aware of the treatment assignment, which could introduce bias, though outcome assessment was blinded.
- The study had broad inclusion criteria regarding ICH location and size, potentially including patients with smaller or less critical hemorrhages who might have good outcomes regardless of intervention.
- The study did not evaluate long-term outcomes beyond 90 days for the primary efficacy endpoint.
- The trial's funding by the manufacturer of andexanet alfa (AstraZeneca) could be a potential source of bias, despite statements of independence in study conduct.
Citation
N Engl J Med 2024;390:107-17. DOI: 10.1056/NEJMoa2310234