Background
Phase 3 double-blind RCT (MAGELLAN). 8,101 acutely ill medical patients hospitalized for ≥3 days with reduced mobility, aged ≥40, and at elevated VTE risk (modified IMPROVE VTE score ≥2 or Padua score ≥4). Compared rivaroxaban extended prophylaxis (35 days) to standard enoxaparin prophylaxis (10 days) followed by placebo. Extended-duration prophylaxis was driven by evidence that VTE risk persists beyond standard 6–14 day hospital-based prophylaxis in medical patients.
Interventions and follow up
Arm A: Rivaroxaban 10 mg once daily × 35 day
Arm B: Enoxaparin 40 mg SC once daily × 10 days, then placebo × 25 day
Primary endpoint: Composite of asymptomatic proximal DVT, symptomatic DVT or PE, or VTE-related death; assessed at day 10 (non-inferiority) and day 35 (superiority)
mFollow up: 35 days (primary); 90 day
Arm B: Enoxaparin 40 mg SC once daily × 10 days, then placebo × 25 day
Primary endpoint: Composite of asymptomatic proximal DVT, symptomatic DVT or PE, or VTE-related death; assessed at day 10 (non-inferiority) and day 35 (superiority)
mFollow up: 35 days (primary); 90 day
Results
VTE composite at day 10: 2.7% (rivaroxaban) vs 2.7% (enoxaparin); RR 0.97 — non-inferior (P=.0025)
VTE composite at day 35: 4.4% (rivaroxaban) vs 5.7% (enoxaparin/placebo); RR 0.77, 95% CI 0.62–0.96 — P=.02 (superior)
Major bleeding at day 10: 0.6% vs 0.3%; RR 2.32, 95% CI 1.21–4.44 — P=.009
Major bleeding at day 35: 1.1% vs 0.4%; RR 2.50, 95% CI 1.46–4.27 — P=.001
VTE composite at day 35: 4.4% (rivaroxaban) vs 5.7% (enoxaparin/placebo); RR 0.77, 95% CI 0.62–0.96 — P=.02 (superior)
Major bleeding at day 10: 0.6% vs 0.3%; RR 2.32, 95% CI 1.21–4.44 — P=.009
Major bleeding at day 35: 1.1% vs 0.4%; RR 2.50, 95% CI 1.46–4.27 — P=.001
Adverse events
Main adverse events: Major bleeding significantly higher with rivaroxaban at both day 10 (0.6% vs 0.3%; P=.009) and day 35 (1.1% vs 0.4%; P=.001). Clinically relevant non-major bleeding also increased (2.8% vs 1.2%; P<.001). Fatal bleeding: 0.1% each arm. Net clinical outcome (VTE + major bleeding) was unfavorable for rivaroxaban. The VTE benefit was offset by the bleeding excess.
Conclusions
While rivaroxaban extended prophylaxis reduced VTE at 35 days, it significantly increased major and clinically relevant bleeding, resulting in an unfavorable risk-benefit balance for unselected acutely ill medical patients.
Key Limitations
Key Limitations: Asymptomatic proximal DVT (ultrasound-detected) drives much of the composite endpoint — clinical relevance debatable. Unselected high-risk medical patients include a heterogeneous group with varying bleeding risk. The bleeding excess rendered the net clinical benefit negative. Subsequently, the APEX trial (betrixaban) and MARINER trial (rivaroxaban 10 mg in higher-risk patients) attempted patient selection enrichment. Extended prophylaxis has not been broadly adopted for unselected acutely ill medical patients due to this unfavorable net benefit.
Clinical Context
MAGELLAN failed to establish extended rivaroxaban prophylaxis as a standard for unselected medical inpatients. The bleeding excess (particularly GI and urinary) in an already-frail population was prohibitive. Subsequent trials (APEX, MARINER) and meta-analyses identified that only high-risk/low-bleeding-risk patients may benefit. Current CHEST and ACCP guidelines do not recommend extended pharmacological prophylaxis for routine medical inpatients; risk stratification using validated scores (IMPROVE Bleed, IMPROVE VTE) is key for any extended prophylaxis decision.
References