Background: Hematopoietic stem cell transplantation (HSCT), the only available curative therapy for sickle cell disease (SCD), remains hampered by the lack of histocompatible stem cell donors. Most patients will not have a suitable human leukocyte antigen(HLA) matched sibling donor. In addition, SCD affects ethnic groups that are underrepresented in stem cell donor registries worldwide.

Objective: to assess the probability of having a potential allelic HLA matched unrelated donor(MUD) for HLA-A,B and DRB1 loci (6/6) in international donor registries for patients with SCD.

Methods: 185 patients with SCD were included, 116 from Brazil, of whom 23 underwent HSCT, and 69 patients who underwent HSCT in centers reporting to the European Society for Blood and Marrow Transplantation (EBMT). All patients had HLA typing available at intermediate or high resolution. For intermediate resolution, using the National Marrow Donor Program (NMDP) code, we assigned alleles based on allele frequencies.We performed HLA haplotype estimation using the HaploStats website, which describes HLA haplotype frequency from the NMDP registry for the following ethnic groups: Caucasian, African-American, Asian, Hispanic and Native American. Because Hispanic is not a primary ethnicity, we did not consider this group in our analyses. Based on haplotype frequency of each ethnic group, we defined the most likely ethnicity for each estimated haplotype; those with frequency >1:1000 in all ethnic groups were named common. Unrelated donor search was done using the World Marrow Donor Association (WMDA) algorithm, which is based on haplotype matching. A potential allelic donor was defined as a full match high resolution 6/6 donor. Because it is described that testing at least 5 potential allelic donors simultaneously increases the chances of having a real donor, we assessed the probability of finding at least 1 and at least 5 potential allelic donors. Patients who received HSCT from MUD were excluded from donor searches (n=10).Comparisons of probabilities of having potential allelic donors between Brazilian and EBMT cohorts were performed by chi-square.

Results: In the Brazilian cohort, from 181 HLA haplotypes, 45% were classified as African-American, 18% common, 12% Caucasian, 9% Amerindian and 16% could not be classified. In the EBMT cohort, from 116 HLA haplotypes, 70% were classified as African-American, 8% common, 6% Caucasian, 3% Amerindian and 13% were not classified. Although Brazilians showed greater genetic admixture, chances of finding at least one potential allelic MUD were 47% in both groups (p-value not significant) and chances of having at least 5 potential allelic MUD were 24% for Brazilians and 15% for EBMT (p-value not significant). Overall, most potential allelic MUD were found in the NMDP registry, followed by the Brazilian registry (REDOME) and by the German registry (ZKMD); for the Brazilian cohort, most potential allelic MUD were found in REDOME.

Discussion: Migration from Africa to Brazil started at the colonial period, and interethnic admixture have been occurring since then, explaining the higher diversity observed in the Brazilian cohort. Despite differences in ethnic composition, chances of having at least one potential allelic MUD are identical, probably because carrying at least one African or Amerindian haplotype decreases the chances of a full HLA matching. Although we demonstrated higher probabilities of finding a potential allelic MUD in SCD than previous studies, the chances are still low, therefore further strategies are required to increase donor representativeness for SCD. In this setting, alternative sources, such as haploidentical HSCT and cord blood, should be considered. Also, our study might help to predict the probabilities of finding a MUD for patients with SCD. This is important because given that HSCT in SCD has better results if performed at earlier age, knowing which patients are less likely to find a MUD might influence therapy management.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.