Intravenous (IV) busulfan (BU) gives more predictable exposure than the oral form and is easier to administer once daily than the conventional four times daily schedule. Fludarabine is immunosuppressive, antileukaemic and perhaps better tolerated than cyclophosphamide. Antithymocyte globulin may reduce graft-versus-host disease (GVHD). We have investigated a conditioning regimen combining these agents in AML and are exploring the possibility that adding a low dose of total body irradiation (TBI) may enhance cytoreduction and decrease relapse rate without increasing toxicity. Of a total of 133 pts 86 (65%) had good-risk (GR) leukaemia in CR1 (n=64) or CR2 (n=22), 47(35%) had more advanced disease (HR). Of the 120 instances where cytogenetics were available 13 (11%) had good risk karyotypes while 84 (70%) of karyotpyes were standard risk and 23 (19%) high risk. Donors were matched siblings (MRD) for 66 (50%) and alternate donors (AD, 55 unrelated and 11 mismatched related) for 67 (50%). Cell source was blood in 64 MRD (98%) and 40 (60%) AD pts. Median pt age was 43 (range 16–65) years. All pts received fludarabine 50mg/m2 on days -6 to -2 and IV BU (Busulfex, ESP Pharma) at a “myeloablative” dose of 3.2 mg/kg daily days -5 to -2 inclusive. Additional TBI 200cGy x 2 was given to 43 pts (31 GR, 12 HR, 16 MRD, 27 AD). The distribution of cytogenetic changes was not significantly different in the TBI and no TBI groups (good risk 13% vs. 10%, standard risk 65% vs. 73% and high risk 23% vs. 14%). All pts received cyclosporine A, “short course” methotrexate with folinic acid and Thymoglobulin (Genzyme) 4.5 mg/kg in divided doses over 3 consecutive days pretransplant finishing D0. Follow-up of survivors is 4-71 months, median 30.5. Acute GVHD grade II-IV occurred in more recipients of SCT from AD (27±6% vs 11±4%, p = 0.02). Incidence of grade III-IV acute GVHD at 9±4% and 8±4% and chronic GVHD (at 2 years) at 59±7% and 62±10% was similar. Projected TRM at 3 years was 10±4% for MRD recipients vs 16±5% for the AD group (p = ns) and was 12±6% vs 15±4% with and without TBI respectively (p = ns). Relapse at 3 years was not significantly affected by donor (GR MRD 27±7% vs AD 21±7%, HR MRD 83±10% vs AD 68±15%) or by acute or chronic GVHD. There was less relapse after the TBI containing regimen in HR recipients (35±17% vs 81±8% at 3 years, p = 0.015), the difference for GR pts at 10±5% vs 31±7% respectively did not reach significance. Projected outcomes at 3 years for GR pts are OS 83±8% with TBI vs 66±6% without (p = ns), and DFS 83±8% with vs 60±7% without TBI respectively (p = 0.03). For HR disease projected OS was 61±15% with TBI and 28±8% without TBI (p = 0.06), DFS was 48±15% and 15±6% respectively (p = 0.06). For GR pts OS is estimated to be 71±8% and 68±8% and DFS 68±7% and 67±8% with MRD and AD respectively. Corresponding figures for HR disease are OS 29±10% 36±11% and DFS 13±8% and 26±13%. None of these differences are significant. This regimen appears relatively well tolerated and gives equivalent results for MRD and AD. Further study is needed to confirm that additional TBI does not add to TRM but may reduce relapse.