Abstract 835

Recently, Stevens et al reported that unrelated cord blood (UCB) search algorithms should be modified including the identification of the HLA mismatches direction given priority to graft-versus-host mismatch direction (GvH). In this study, 1202 patients were analyzed, 870 with malignant disorders, after a single UCB transplantation (UCBT) in US Centers from 1993 to 2006. Nearly two-thirds were transplanted for leukemia (ALL, AML or CML) and, among these, 28% had advanced disease status. Based in HLA mismatches direction, 890 had bidirectional mismatch (Bidir), 58 had GvH, 40 had host-versus-graft mismatch direction (HvG) and 69 had fully matched grafts. The authors considered the GvH and HvG patients subgroups with either one (5/6) or two (4/6) HLA mismatches and found that patients with GvH had better engraftment, lower non-relapse mortality (NRM), overall survival (OS) and treatment failure (Blood 2011). In order to analyze the association of HLA mismatch direction on outcomes after UCBT in other cohort of patients, we performed a retrospective analysis on 1996 patients receiving a single CB transplant in Eurocord centers between 1994 and 2009. All patients received a single unit, UCBT as a first allogeneic transplant for malignant or non-malignant diseases, with 0, 1 or 2 out of 6 HLA mismatches (HLA-A and HLA-B at the antigen level and HLA-DRB1at the allelic level). Patients were classified in GvH direction (when donor was homozygous at an HLA locus but, the patient had two antigens identified [one matching the donor] at that locus), host-versus-graft direction (when the patient was homozygous at a locus but, the donor had two antigens identified [one matching the donor] at that locus) and bidirectional mismatch (mismatched HLA antigen presented in both recipient and donor). Those groups were analyzed separately for 0–1 HLA and 2 HLA mismatches groups. The median age was 10 years (0.06 - 65 years) and median weight 33Kg at time of UCBT. Almost 79% of patients were transplanted for malignant diseases (AML, ALL, MDS, CML and others) and 21% for non-malignant diseases. Twenty two percent of patients received reduced intensity conditioning with an infused CD34+ dose >1.6 × 105/Kg and infused TNC dose >3.8 × 107/Kg of recipient body weight. Two hundred and sixty five patients (13%) were HLA identical, 853 (43%) had 1 HLA difference and 878 (44%) had 2 HLA differences. The overall HLA mismatch direction was: 734 patients had one and 716 had two bidirectional mismatches, 66 had GvH direction (63: 5/6; 3: 4/6) and 62 had HvG direction (56: 5/6; 6: 4/6) mismatches. The remaining 153 pairs had other various combinations of HLA mismatch directions. The median follow-up time was 36 months and the median times for neutrophil and platelet recoveries were 24 and 42 days, respectively. Cumulative incidence (CI) of ANC recovery was 78%; we found a better neutrophil recovery in the group of patients given a 4/6 graft with a GvH mismatch direction (HR: 1.3 p=.01). At day +100, CI of grade II–IV acute GvHD was 33% (8% grade III, 7% grade IV), and it was not associated with any HLA mismatch direction. At 1-year, CI of Non Relapse Mortality (NRM) was 25%. Estimated 3 y-DFS was 38%. Relapse, NRM and DFS were not associated with any HLA mismatch direction in multivariate models. When selecting a more homogenous population (n=917) with hematological malignancies transplanted in early and intermediate disease status, receiving a single unit UCBT after myeloablative conditioning between 2000–2009, CI of relapse at 3 years was 32% (GvH: 28%, HvG: 30% and Bidir: 29%) and 2 y-DFS 34%. Also, no association between HLA mismatch direction and DFS was observed in a multivariate analysis. Based on this data, we showed that in a small group of patients given a 4/6 graft, improved neutrophil recovery seemed to be associated with the HLA mismatch in the GvH direction. Importantly, we did not show any evidence that HLA mismatch direction had any significant impact on outcomes after UCBT as it has been recently shown. Therefore, based in this larger series of patients, we do not recommend selecting a cord blood unit based on the direction of the HLA mismatch.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution