Busulfan is an alkylating agent that is frequently used as the ablative component of allogeneic stem cell transplants in conjunction with immunosuppressive drugs such as cyclophosphamide or fludarabine. Busulfan has recently become available for IV use and has been shown to be effective and less toxic due to more reliable pharmacology than the oral form of this drug. While prolonged administration of this drug has not been feasible in the past, the parenteral formulation now makes this possible and permits re-evaluation of the efficacy, toxicity, and optimal dosing of this compound. Such an approach may reduce the dose-limiting hepatic, pulmonary and CNS toxicities that are seen with this compound. Here we report on 15 patients treated on a study using a continuous 90 hour infusion of this drug at 12.8 mg/kg adjusted body wt along with 30 mg/m2 fludarabine qd x 5 and alemtuzamab at doses of 0, 30, or 90 mg depending on the disease (myeloid or non-myeloid) and donor (MRD or MUD) followed by SCT. GHVD prophylaxis also included tacrolimus for all pts and short-course methotrexate for MRD patients who did not receive alemtuzamab. Subjects included 11 males and 4 females, median age 38, 6 MUD, 9 MRD, and 9 AML, 2 APL, 2 CML, 1 MDS and 1 ALL pt. All were high-risk based on being in 2nd remission or beyond (4), or having MDS or trilineage dysplasia(4), poor cytogenetics (3), refractory disease (2), prior transplant (1), or advanced CML (1). There have been no grade 3 or 4 toxicities due to drug infusion or during the initial period of aplasia. Neutrophil and platelet engraftment occurred on day + 14 and day + 16 respectively. There was one graft failure in a 54 yo man with MDS/AML who received a sex mismatched MUD transplant and subsequently underwent successful re-transplant on day 50. There have been two cases of grade II, one grade III and one grade IV aGVHD in 2 MRD and 2 MUD pts. Relapses have occurred in 5 pts after a median f/u of 10 months. Pharmacokinetics were obtained on 11 pts and compared the AUC obtained with a test dose given the day prior to initiation of the 90 hour infusion and the AUC of the continuous infusion. Test dose AUC predicted the full dose AUC with > 90% accuracy (range 3–19%, precision and bias of 9.5 and 1.4%). Mean daily AUC was 3539 umol min (range 1936 to 5591) with minimal variation between day 1 and 4 for the full-dose infusion. This is similar to the mean AUC observed in a control group of 8 pts treated with q 6 hr dosing and reflects the 3-fold inter-patient variability that has been reported when the parenteral drug is administered by intermittent infusion. 3 of the 4 pts with AUC data who have relapsed have had AUC concentrations below the mean. No AUC targeting was undertaken in this study but the predictive capacity of the test-dose strategy and the low AUC suggest that targeted dosing would be feasible and desirable with this approach. Studies with targeted dosing and prolonged infusion are currently underway to identify better the risks, toxicities and MTD of this approach.