The purpose of this study was to evaluate the influence of busulfan (BU) pharmacokinetics on survival, grades II-IV acute graft-versus-host disease (GVHD), non-relapse mortality (NRM) and relapse in a group composed of 45 children (<18 years) and seven adults with acute myeloid leukemia (AML) in first complete remission and undergoing hematopoietic stem cell transplantation (SCT). Fifty-two patients underwent autologous (n = 25) or allogeneic (n = 27) SCT. The median age was 8.9 years (range 0.6-53 years). Conditioning therapy consisted of BU and cyclophosphamide. Improved disease-free survival was found in those patients with a steady-state concentration of BU (CssBU) below the median (<578 mg/ml, P = 0.05), and the same trend was noted for overall survival (P = 0.07). This was secondary to a higher incidence of NRM in the group of patients with CssBU above the median (P = 0.06). There was no significant correlation with CssBU and relapse (P = 0.31). No association between CssBU and GVHD was found in allogeneic patients (P = 0.30). Relapse was evaluated among the subgroups of age (< or >10 years) and transplant type (allogeneic or autologous) with no statistically significant association observed among these factors. Multiple regression analysis for relapse revealed no significant correlation with CssBU above or below the median, age, or transplant type. In this study, CssBU below the median did not correlate with an inferior outcome for patients with AML. Pharmacokinetic dosing of BU may be important for prevention of NRM but does not appear to influence the risk of relapse in this largely pediatric population with AML.