A B S T R A C TWe previously reported in a French prospective randomized study that transplantation of 2 unrelated cord blood (UCB) units instead of 1 unit does not decrease the risk of transplantation failure but may enhance alloreactivity. Here we evaluated the influence of pretransplantation minimal residual disease (MRD) on leukemia relapse and survival after single-versus double-UCB transplantation (UCBT). Among 137 children and young adults who underwent UCBT in this randomized study, 115 had available data on MRD assessment done immediately before initiation of the pretransplantation conditioning regimen. MRD was considered positive at a level of 10 ¡4 , which was the case of 43 out of 115 patients. Overall, the mean 3-year survival probability was 69.1 § 4.4%, and it was not significantly influenced by the MRD level: 70.7 § 5.4% in MRD-negative (<10 ¡4 ) patients (n = 72), 71.1 § 9.4% in MRD-positive patients with 10 ¡4 MRD <10 ¡3 (n = 26) and 58.8 § 11.9% in MRD-positive patients with 10 ¡3 (n = 17). In the MRD-positive group, the mean risk of relapse was significantly lower in the double-UCBT arm compared with the single-UCBT arm (10.5 § 7.2% versus 41.7 § 10.4%; P = .025) leading to a higher mean 3-year survival rate (82.6 § 9.3% versus 53.6 § 10.3%; P = .031). This difference was observed only in patients who had not received antithymocyte globulin during their conditioning regimen. In the MRD-negative group, there was no difference between the single-and the double-UCBT arms. We conclude that even in cases of positive pretransplantation MRD, UCBT in children and young adults with acute leukemia yields a high cure rate, and that a double-unit strategy may enhance the graft-versus-leukemia effect and survival in these patients.