In vivo alemtuzumab reduces the risk of graft-versus-host disease (GVHD) and nonrelapse mortality after reduced intensity allogeneic transplantation. However, it also delays immune reconstitution, leading to frequent infections and potential loss of graft-versus-tumor responses. Here, we tested the feasibility of alemtuzumab dose deescalation in the context of fludarabine-melphalan conditioning and human leukocyte antigen (HLA)-identical sibling transplantation. Alemtuzumab was given 1-2 days before graft infusion, and dose reduced from 60 mg to 20 mg in 4 sequential cohorts (total n = 106). Pharmacokinetic studies were fitted to a linear, 2-compartment model in which dose reduction led to incomplete saturation of CD52 binding sites and greater antibody clearance. Increased elimination was particularly evident in the 20-mg group in patients who had CD52-expressing tumors at time of transplantation. The 20-mg dose was also associated with greater risk of severe GVHD (acute grade III-IV or chronic extensive) compared with > 20 mg (hazard ratio, 6.7; 95% CI, 2.5-18.3). In contrast, dose reduction to 30 mg on day -1 was associated with equivalent clinical outcomes to higher doses but better lymphocyte recovery at 12 months. In conclusion, alemtuzumab dose reduction to 30 mg is safe in the context of reduced intensity conditioning and HLA-identical sibling transplantation. This trial was registered at http://www.ncrn.org.uk as UKCRN study 1415.