Finding an HLA-matched donor for hematopoietic stem cell transplantation — from family typing to registry search to alternative sources
When a patient needs a hematopoietic stem cell transplant (HSCT), the search for a donor begins at home. Full siblings share both parents, and HLA genes are inherited as a linked block (haplotype) from each parent — making a full sibling the statistically fastest and historically most reliable route to a matched donor.
HLA genes cluster tightly on chromosome 6 and are inherited together as a single block, called a haplotype, from each parent. A child therefore receives one haplotype from mother and one from father — four possible combinations exist among full siblings (maternal-A/paternal-A, maternal-A/paternal-B, maternal-B/paternal-A, maternal-B/paternal-B).
Because each combination is equally likely, any two full siblings have a 25% chance of inheriting the exact same haplotype pair — a full match. This is dramatically higher than the probability of matching an unrelated stranger, whose HLA genotype is essentially random relative to the patient's. Parents and children, by contrast, always share exactly one haplotype (never both), making them obligate haploidentical — a half-match, not a full match.
With one sibling, the chance of a match is 25%. With four full siblings tested, the cumulative chance of finding at least one full match rises to roughly 68% — which is why centers typically type all available full siblings before moving to an unrelated search.
Family typing uses a blood or buccal swab sample from the patient and each available full sibling, processed by high-resolution molecular HLA typing (typically next-generation sequencing or sequence-specific oligonucleotide/primer methods). The key loci examined are HLA-A, -B, -C, and -DRB1, with many centers also typing HLA-DQB1.
Results are compared allele-by-allele. Because each locus carries two alleles (one per haplotype), a full match at the four core loci is described as an "8 of 8" or "8/8" match. Family typing is fast — results are typically available within one to two weeks — and inexpensive relative to registry-based searches, which is why it is always the recommended starting point when the clinical timeline allows.
If no full sibling is available or none match, some centers extend testing to half-siblings (lower match probability, since only one parent is shared) or, in rare cases, more distant relatives — though these searches offer diminishing statistical returns compared to a well-resourced unrelated registry search.
In parallel with family typing, most transplant centers simultaneously initiate a preliminary unrelated registry search, so that if family typing does not yield a match, the registry process has already had a head start rather than beginning from zero.
Not every available donor is equally suitable. Transplant centers grade potential donors by how many of the key HLA loci match at high resolution, producing a match grade that directly informs expected outcomes — from graft-versus-host disease (GVHD) risk to graft rejection and relapse probability.
A "matched" donor is defined by concordance at the classical HLA Class I loci — HLA-A, -B, and -C — and the Class II locus HLA-DRB1. Because each person carries two alleles per locus (one inherited from each parent), matching these four loci means comparing 8 total alleles, hence the familiar "8/8" shorthand.
Many transplant programs now also type HLA-DQB1, extending the panel to 10 alleles ("10/10"), since mismatches at this locus have been shown in some studies to modestly affect outcomes, particularly for unrelated donor transplants.
Donors are ranked along a descending scale of suitability:
• 8/8 (or 10/10) matched — all core loci match at high resolution; the reference standard, associated with the lowest rates of severe GVHD and graft rejection • 7/8 — a single-allele mismatch at one locus; still usable, with somewhat elevated GVHD and non-relapse mortality risk depending on which locus is mismatched • 6/8 and below — two or more mismatches; generally reserved for situations where no better-matched donor is available, given progressively higher immunologic risk
Not all mismatches carry equal weight — a mismatch at HLA-C or HLA-DRB1 tends to carry more clinical risk than certain other single-locus mismatches, so the position of the mismatch matters as much as the count.
A single-allele mismatch (7/8) is not an automatic disqualifier — it is a risk-adjustment factor weighed against disease urgency, donor age and health, and availability of better alternatives.
Early HLA typing methods (serologic or "antigen-level" typing) could only distinguish broad antigen groups, potentially missing clinically important differences between closely related alleles. Modern high-resolution ("allele-level") molecular typing — most commonly next-generation sequencing — resolves HLA genes down to the specific allele, revealing mismatches that antigen-level typing would have missed.
Because a donor who appears matched at low resolution can turn out to be mismatched at high resolution, current standards require high-resolution confirmatory typing of both patient and donor before a transplant proceeds, regardless of the resolution used during initial registry searching.
When no matched family donor exists, the search expands to international unrelated donor registries — vast databases of HLA-typed volunteers linked through a global network. Search success, however, is far from uniform: it depends heavily on how well the patient's genetic background is represented in the pool of registered volunteers.
A blood or swab sample from the patient generates a high-resolution HLA profile, which is submitted to national donor registries (such as the NMDP/Be The Match in the US) and, through the World Marrow Donor Association (WMDA), cross-searched against affiliated registries worldwide. A preliminary search returns a ranked list of potential matches based on the typing already on file for each volunteer.
The strongest candidates are then contacted for confirmatory high-resolution typing, since many registry entries were typed at lower resolution when the volunteer originally joined. This confirmatory step can eliminate apparent matches that turn out, on closer inspection, to be mismatched at the allele level.
HLA genes are extraordinarily polymorphic, and the frequency of specific haplotypes varies substantially across populations. Because most large registries were built predominantly from volunteers of European ancestry, patients of European descent tend to have the highest likelihood of finding an 8/8 unrelated match.
Patients from populations that are historically underrepresented in donor registries — including many patients of African, mixed-race, and some Asian, Hispanic, and Indigenous backgrounds — often face a substantially lower probability of finding a fully matched unrelated donor, sometimes below 30%. This disparity is a major, ongoing driver of donor-recruitment campaigns and of increased reliance on alternative donor sources such as haploidentical transplant and cord blood.
Ancestry-driven disparities in registry match rates are a primary reason alternative donor strategies (haploidentical transplant, cord blood) have become mainstream rather than last-resort options for many patients.
Once a promising unrelated donor is identified, the process continues through several sequential steps: confirmatory high-resolution typing, a donor health history questionnaire, a physical examination, and infectious disease screening. Coordinating these steps across registries, time zones, and collection centers takes time.
End-to-end, an unrelated donor search — from initial query to a donor cleared and scheduled for collection — commonly takes several weeks to a few months, considerably longer than family typing, which is why it typically proceeds in parallel with, rather than strictly after, family testing whenever the clinical situation allows.
For patients without a matched family or unrelated donor, three alternative sources have transformed transplant access: haploidentical (half-matched) relatives, banked umbilical cord blood, and mismatched unrelated donors. Each trades some degree of HLA mismatch for dramatically faster and more reliable donor availability.
A haploidentical donor — typically a parent, child, or sibling — shares exactly one HLA haplotype with the patient, an inevitable consequence of biological inheritance. Because nearly every patient has at least one living haploidentical relative, this source offers near-universal, rapid availability, often identifiable within days.
Historically, the immunologic mismatch made haploidentical transplant high-risk for severe GVHD and rejection. The introduction of post-transplant cyclophosphamide (PTCy) — a short course of chemotherapy given days after the stem cell infusion to selectively eliminate the most alloreactive, rapidly dividing donor and recipient T-cells — has transformed haploidentical transplant from a last resort into a mainstream option with outcomes approaching those of matched donor transplants at many centers.
Public cord blood banks store hundreds of thousands of units collected at birth and HLA-typed in advance, allowing near-immediate availability once a suitable unit is identified — often faster than an adult unrelated donor. Cord blood also tolerates a greater degree of HLA mismatch than adult donor sources, because the immunologically naive cells carried in cord blood are less prone to triggering severe GVHD.
The major limitation is cell dose: a single cord blood unit contains far fewer stem cells than a bone marrow or peripheral blood collection from an adult donor, which can slow engraftment and increase infection risk, particularly in larger patients. Some centers address this by infusing two partially matched cord blood units ("double cord") to boost total cell dose.
When registry searching identifies only a 7/8 or lower-matched unrelated donor, this donor can still be used, generally alongside more intensive immunosuppression (increasingly PTCy-based protocols, mirroring haploidentical approaches) to control the elevated risk of GVHD and rejection associated with the mismatch.
The choice among haploidentical, cord blood, and mismatched unrelated sources depends on donor availability, patient age and fitness, disease type and urgency, and each transplant center's specific expertise and protocol experience — there is no single universally "best" alternative; the decision is individualized.
A thorough donor search takes time — and time is the one resource many transplant candidates do not have in abundance. Balancing the pursuit of the best possible match against the clock of disease progression is one of the most consequential judgment calls in transplant medicine.
Each donor pathway carries a distinct timeline. Family typing generally returns results within one to two weeks. An unrelated registry search — from preliminary query through confirmatory typing, health screening, and collection scheduling — typically requires several weeks to a few months. Identifying and clearing a haploidentical family donor can often be accomplished within days to two weeks, since the donor pool (immediate relatives) is already known and typically local.
Because these timelines run in parallel rather than strictly sequentially at most centers, disease risk stratification — done at diagnosis — determines how much time can be allotted to registry searching before an alternative donor decision must be made.
For patients with slower-progressing disease or those in a stable remission, it is often reasonable to allow the full registry search process to run its course in pursuit of the best-matched donor, since a fully matched donor is generally associated with the lowest risk of GVHD, rejection, and non-relapse mortality.
For patients with rapidly progressing or high-relapse-risk disease, however, waiting months for a registry search to complete can itself become the dominant risk — disease relapse or progression during the search may foreclose the option of transplant altogether. In these situations, proceeding promptly with the best available alternative donor (haploidentical, cord blood, or mismatched unrelated) is frequently the safer overall strategy, even though it trades some immunologic risk for speed.
The central clinical judgment is not "which donor is theoretically best" but "which donor can be transplanted in time to matter" — a calculation that shifts case by case with disease biology, patient status, and how fast a fully matched donor can realistically be secured.
As outcomes with PTCy-based haploidentical transplant and improved cord blood protocols have increasingly approached those of matched donor transplant at experienced centers, the historical pressure to wait indefinitely for a perfect match has eased considerably. Alternative donors are no longer viewed purely as a fallback — they are increasingly a deliberate first-line choice when speed, donor availability, and center expertise favor them.
This shift has been especially significant for patients from populations underrepresented in unrelated donor registries, for whom alternative donor pathways now offer a much more reliable and timely route to transplant than registry searching alone ever could.