Background
Phase III, open-label RCT (BMT CTN 0901). 272 patients aged 18–65 with AML in first or second CR, or MDS (IPSS intermediate-2 or high risk), eligible for both myeloablative (MAC) and reduced-intensity (RIC) allogeneic HCT. Question: does MAC improve long-term disease control vs RIC despite increased NRM risk? First prospective randomized comparison.
Interventions and follow up
Arm A (MAC): busulfan 3.2mg/kg/day IV ×4 days + cyclophosphamide 60mg/kg/day ×2 (Bu/Cy), or TBI 12–13.2 Gy + cyclophosphamide 60mg/kg ×2 (TBI/Cy)
Arm B (RIC): fludarabine 30mg/m² ×5 days + busulfan 0.8mg/kg IV q6h ×4 doses, or fludarabine + melphalan 140mg/m²
GVHD prophylaxis: tacrolimus + methotrexate (MRD) or tacrolimus + mycophenolate (MUD)
Primary endpoint: Overall survival at 18 months
Median follow up: 38.7 months
Arm B (RIC): fludarabine 30mg/m² ×5 days + busulfan 0.8mg/kg IV q6h ×4 doses, or fludarabine + melphalan 140mg/m²
GVHD prophylaxis: tacrolimus + methotrexate (MRD) or tacrolimus + mycophenolate (MUD)
Primary endpoint: Overall survival at 18 months
Median follow up: 38.7 months
Results
18-month OS: 47.9% (MAC) vs 25.4% (RIC) — HR 0.60, 95% CI 0.40–0.90, P=.003
3-year relapse: 44.9% (MAC) vs 65.3% (RIC) — P<.001
3-year NRM: 15.0% (MAC) vs 11.9% (RIC) — P=.41 (NS)
3-year OS: 35.8% (MAC) vs 15.8% (RIC) — P=.004
3-year relapse: 44.9% (MAC) vs 65.3% (RIC) — P<.001
3-year NRM: 15.0% (MAC) vs 11.9% (RIC) — P=.41 (NS)
3-year OS: 35.8% (MAC) vs 15.8% (RIC) — P=.004
Adverse events
GVHD: grade ≥2 acute GVHD 44.5% (MAC) vs 40.2% (RIC), NS; grade ≥3 acute GVHD 16.5% vs 15.5%, NS; chronic GVHD 53.6% vs 52.1%, NS
Infections: grade ≥3 infections 58.3% (MAC) vs 52.1% (RIC), NS
Non-relapse mortality: similar between arms, largely offsetting the relapse benefit of MAC
Infections: grade ≥3 infections 58.3% (MAC) vs 52.1% (RIC), NS
Non-relapse mortality: similar between arms, largely offsetting the relapse benefit of MAC
Conclusions
MAC provided significantly superior OS and lower relapse vs RIC in fit AML/MDS patients, driven by dramatically lower relapse (44.9% vs 65.3%). NRM did not differ significantly, overturning the assumption that RIC's lower NRM would offset higher relapse.
Key Limitations
Terminated early (planned N=356, enrolled 272) when interim analysis showed OS superiority for MAC, potentially inflating effect size. Eligibility required suitability for MAC (≤65, good organ function), a fit subpopulation not generalizable to unfit/older patients. Regimen heterogeneity within arms (Bu/Cy vs TBI/Cy; Flu/Bu vs Flu/Mel) and disease heterogeneity (AML CR1, CR2, MDS) limit mechanistic and subgroup conclusions.
Clinical Context
BMT CTN 0901 altered practice: for fit patients (≤65, good PS) with AML/MDS eligible for either conditioning intensity, MAC is preferred at most centers, while RIC remains standard for those ineligible (older age, comorbidities). Supports the principle that disease control rather than NRM reduction drives OS in MAC-fit patients. Integrates with ELN 2022 and ASTCT transplant-timing recommendations.