In absence of available matched-related or unrelated donor (MUD), mismatched unrelated donors (MMUD) and unrelated cord blood (UCB) are both considered to be suitable donors, with similar post-transplant overall survival. In most of these retrospective comparisons, HLA typing of adult donors was performed at eight loci. The aim of this study was to compare the outcome of patients transplanted from UCB (N 5 64) with those transplanted from 9/10-HLA MMUD (N 5 84) or 10/10-HLA MUD (N 5 196). In multivariate analysis, UCB was associated with less Grade II-IV acute GVHD in comparison with MUD (aHR 1.97, 95% CI 1.19-3.27, P 5 0.009) and MMUD transplants (aHR 1.79, 95% CI 1.02-3.15, P 5 0.042), while the cumulative incidence of chronic GVHD was not significantly different between the three groups. Overall survival (OS), non-relapse mortality, and relapse were not different between MMUD and UCB transplantation, whereas OS was impaired after UCB in comparison with MUD (aHR 0.65, 95% CI 0.43-0.99, P 5 0.043). Factors also impacting OS were the donor/recipient CMV serostatus (Donor2/Recipient1 aHR 1.76, 95% CI 1.23-2.52, P 5 0.002 compared with D2/R2), the donor/recipient gender combination (Female/Male versus other combinations aHR 1.57, 95% CI 1.11-2.22, P 5 0.012) and disease risk (aHR 1.58, 95% CI 1.05-2.38, P 5 0.027 for high vs. low risk disease). Our data confirm that UCB and 9/10-HLA MMUD are both relevant alternative options when no 10/10-HLA donor is available. Donor/recipient gender combination and CMV serostatus had a significant impact on survival and may be taken into account, along with donor type, in the setting of MMUD and UCB transplants.