The Registry of Unmet Need: A World Marrow Donor Association Analysis of Patients Without an HLA Match

Abstract

While the World Marrow Donor Association global database currently offers approximately 42.7 million potential donors and cord blood units to patients in need of haematopoietic cell transplant, lack of eight HLA-matched donors remains a significant barrier. The Registry of Unmet Need (RUN) Project seeks to address disparities in transplant access for patients with rare HLA genotypes, particularly those from populations that have been historically underrepresented and underserved by global donor registries. Patients eligible for this study searched for an unrelated donor for transplant between 2015 and 2017 and, at that time, lacked a potential eight-of-eight HLA-matched unrelated donor (MUD). Sixteen donor registries contributed data from 3654 patients using standardised data-collection project templates. To address this unmet need, pooled data were analysed to identify trends and inform global recruitment strategies. Patient genotypes were queried against the global inventory at later timepoints in 2018 and 2023 to determine whether potential matches had been recruited within the years since the initial search. Patient haplotypes were imputed using an open-source method referencing US population frequencies. The imputation process used five continental reference populations and 21 detailed populations derived from the NMDP database. The method provided a Bayesian inference of population membership. A control group consisting of US patients that yielded 1000 or more potential matches was used for comparison. RUN patient haplotype and genotype frequencies were substantially lower compared with controls; both the more frequent and less frequent haplotypes in RUN patients were found to be approximately 100 times less common than those in the control group. We identified 782 potential cases in which a potential MUD was recruited after the initial RUN patient search was performed; while this result is being further investigated, clear patterns of where these new matches can be found have emerged; typically, new matches are found outside the country where the patient search was initiated. Our findings demonstrate that rare haplotypes are the primary barrier to identifying a MUD; the presence of rare alleles or haplotype combinations, as with multi-race ancestry, is rarely the cause. Although strategic donor recruitment efforts will help improve MUD access, patient transplants should not be delayed in pursuit of a MUD when viable alternative options are available.

Keywords: Bayes theorem; HLA; ethnicity; haematopoietic stem cells; haplotypes; registries.

FIGURE 1

Maps of the global distribution of (A) the countries submitting patients to the RUN cohort, (B) the RUN patient cohort country of origin, (C) the WMDA donors and (D) the global population.

FIGURE 2

A three‐level Sankey diagram of (left) the regional origin of the RUN patients, (middle) the self‐assigned or presumed race category and (right) the population category assigned using an HLA‐based classifier (15 patients assigned to multiple population categories are not displayed).

FIGURE 3

Genotype frequencies for the RUN cohort (red) and a control set of US patients with an eight‐of‐eight match (blue). The vertical dashed line represents 1/the world population. The green dotted line represents 1/the WMDA global donor registry population.

FIGURE 4

Haplotype and genotype frequencies for the run patients (red) and a control set of US patients with an eight‐of‐eight match (blue). The top graph shows the frequency for the higher frequency of the two haplotypes of the patient, the middle graph shows the lower frequency haplotype, and the bottom graph is based on the haplotypes with a regression curve for the two patient cohorts. The green dotted lines indicate a threshold of 1 in 1000 (common) and the red dotted line indicates the upper bound of rare starting at 1 in 10,000,000. The grey dotted line is the midpoint of the two extremes at 1 in 100,000.

FIGURE 5

RUN new match rate by donor region and donor registry size. Each point represents all registries from that region and the total number of donors (on the x‐axis) and the rate at which donors from this region provided new matches for RUN patients (on the y‐axis).