HLA diversity in ethnic populations can affect detection of donor-specific antibodies by single antigen beads.

Introduction: In solid-organ transplantation, human leukocyte antigen (HLA) donor-specific antibodies (DSA) are strongly associated with graft rejection, graft loss, and patient death. The predominant tests used for detecting HLA DSA before and after solid-organ transplantation are HLA single antigen bead (SAB) assays. However, SAB assays may not detect antibodies directed against HLA epitopes that are not represented in the SAB. The prevalence and potential impact of unrepresented HLA epitopes are expected to vary by ethnicity, but have not been thoroughly investigated. To address this knowledge gap, HLA allele frequencies from seven ethnic populations were compared with HLA proteins present in SAB products from two manufacturers to determine unrepresented HLA proteins. Materials: Allele frequencies were obtained from the Common, Intermediate, and Well Documented HLA catalog v3.0, and frequencies of unrepresented HLA types were calculated. Next-generation sequencing was used to determine HLA types of 60 deceased solid-organ donors, and results were used to determine if their HLA-A, -B, -C, and -DRB1 proteins were not present in SAB reagents from two vendors. Unrepresented HLA proteins were compared with the most similar protein in SAB assays from either vendor and then visualized using modeling software to assess potential HLA epitopes. Results: For the seven ethnic populations, 0.5% to 11.8% of each population had HLA proteins not included in SAB assays from one vendor. Non-European populations had greater numbers of unrepresented alleles. Among the deceased donors, 26.7% (16/60) had at least one unrepresented HLA-A, -B, -C, or -DRB1 protein. Structural modeling demonstrated that a subset of these had potential HLA epitopes that are solvent accessible amino acid mismatches and are likely to be accessible to B cell receptors. Discussion: In conclusion, SAB assays cannot completely rule out the presence of HLA DSA. HLA epitopes not represented in those assays vary by ethnicity and should not be overlooked, especially in non-European populations. Allele-level HLA typing can help determine the potential for HLA antibodies that could evade detection.

Next generation multiplexing for digital PCR using a novel melt-based hairpin probe design.

Digital PCR (dPCR) is a powerful tool for research and diagnostic applications that require absolute quantification of target molecules or detection of rare events, but the number of nucleic acid targets that can be distinguished within an assay has limited its usefulness. For most dPCR systems, one target is detected per optical channel and the total number of targets is limited by the number of optical channels on the platform. Higher-order multiplexing has the potential to dramatically increase the usefulness of dPCR, especially in scenarios with limited sample. Other potential benefits of multiplexing include lower cost, additional information generated by more probes, and higher throughput. To address this unmet need, we developed a novel melt-based hairpin probe design to provide a robust option for multiplexing digital PCR. A prototype multiplex digital PCR (mdPCR) assay using three melt-based hairpin probes per optical channel in a 16-well microfluidic digital PCR platform accurately distinguished and quantified 12 nucleic acid targets per well. For samples with 10,000 human genome equivalents, the probe-specific ranges for limit of blank were 0.00%-0.13%, and those for analytical limit of detection were 0.00%-0.20%. Inter-laboratory reproducibility was excellent (r 2 = 0.997). Importantly, this novel melt-based hairpin probe design has potential to achieve multiplexing beyond the 12 targets/well of this prototype assay. This easy-to-use mdPCR technology with excellent performance characteristics has the potential to revolutionize the use of digital PCR in research and diagnostic settings.

Donor chimerism and immune reconstitution following haploidentical transplantation in sickle cell disease.

Introduction: We previously reported the initial results of a phase II multicenter transplant trial using haploidentical parental donors for children and aolescents with high-risk sickle cell disease achieving excellent survival with exceptionally low rates of graft-versus-host disease and resolution of sickle cell disease symptoms. To investigate human leukocyte antigen (HLA) sensitization, graft characteristics, donor chimerism, and immune reconstitution in these recipients. Methods: CD34 cells were enriched using the CliniMACS® system with a target dose of 10 x 106 CD34+ cells/kg with a peripheral blood mononuclear cell (PBMNC) addback dose of 2x105 CD3/kg in the final product. Pre-transplant HLA antibodies were characterized. Donor chimerism was monitored 1-24 months post-transplant. Comprehensive assessment of immune reconstitution included lymphocyte subsets, plasma cytokines, complement levels, anti-viral T-cell responses, activation markers, and cytokine production. Infections were monitored. Results: HLA antibodies were detected in 7 of 11 (64%) evaluable patients but rarely were against donor antigens. Myeloid engraftment was rapid (100%) at a median of 9 days. At 30 days, donor chimerism was 93-99% and natural killer cell levels were restored. By 60 days, CD19 B cells were normal. CD8 and CD4 T-cells levels were normal by 279 and 365 days, respectively. Activated CD4 and CD8 T-cells were elevated at 100-365 days post-transplant while naïve cells remained below baseline. Tregs were elevated at 100-270 days post-transplant, returning to baseline levels at one year. At one year, C3 and C4 levels were above baseline and CH50 levels were near baseline. At one year, cytokine levels were not significantly different from baseline. Discussion: These results suggest that haploidentical transplantation with CD34-enriched cells and peripheral blood mononuclear cell addback results in rapid engraftment, sustained donor chimerism and broad-based immune reconstitution.

Cell-free DNA as a solid-organ transplant biomarker: technologies and approaches.

High-quality biomarkers that detect emergent graft damage and/or rejection after solid-organ transplantation offer new opportunities to improve post-transplant monitoring, allow early therapeutic intervention and facilitate personalized patient management. Donor-derived cell-free DNA (DD-cfDNA) is a particularly exciting minimally invasive biomarker because it has the potential to be quantitative, time-sensitive and cost-effective. Increased DD-cfDNA has been associated with graft damage and rejection episodes. Efforts are underway to further improve sensitivity and specificity. This review summarizes the procedures used to process and detect DD-cfDNA, measurement of DD-cfDNA in clinical transplantation, approaches for improving sensitivity and specificity and long-term prospects as a transplant biomarker to supplement traditional organ monitoring and invasive biopsies.

The changing landscape of HLA typing: Understanding how and when HLA typing data can be used with confidence from bench to bedside.

Human leukocyte antigen (HLA) genes are extraordinary for their extreme diversity and widespread impact on human health and disease. More than 30,000 HLA alleles have been officially named and more alleles continue to be discovered at a rapid pace. HLA typing systems which have been developed to detect HLA diversity have advanced rapidly and are revolutionizing our understanding of HLA's clinical importance. However, continuous improvements in knowledge and technology have created challenges for clinicians and scientists. This review explains how differences in HLA typing systems can impact the HLA types that are assigned. The consequences of differences in laboratory testing methods and reference databases are described. The challenges of using HLA types that are not equivalent are illustrated. A fundamental understanding of the continual expansion of our understanding of HLA diversity and limitations in some of the typing data is essential for using typing data appropriately in clinical and research settings.

Stable to improved cardiac and pulmonary function in children with high-risk sickle cell disease following haploidentical stem cell transplantation.

Children with sickle cell disease (SCD) are at high-risk of progressive, chronic pulmonary and cardiac dysfunction. In this prospective multicenter Phase II trial of myeloimmunoablative conditioning followed by haploidentical stem cell transplantation in children with high-risk SCD, 19 patients, 2.0-21.0 years of age, were enrolled with one or more of the following: history of (1) overt stroke; (2) silent stroke; (3) elevated transcranial Doppler velocity; (4) multiple vaso-occlusive crises; and/or (5) two or more acute chest syndromes and received haploidentical transplants from 18 parental donors. Cardiac and pulmonary centralized cores were established. Pulmonary function results were expressed as percent of the median of healthy reference cohorts, matched for age, sex, height and race. At 2 years, pulmonary functions including forced expiratory volume (FEV), FEV1/ forced vital capacity (FVC), total lung capacity (TLC), diffusing capacity of lung for carbon monoxide (DLCO) were stable to improved compared to baseline values. Importantly, specific airway conductance was significantly improved at 2 years (p < 0.004). Left ventricular systolic function (fractional shortening) and tricuspid regurgitant velocity were stable at 2 years. These results demonstrate that haploidentical stem cell transplantation can stabilize or improve cardiopulmonary function in patients with SCD.

Genetics of HLA Peptide Presentation and Impact on Outcomes in HLA-Matched Allogeneic Hematopoietic Cell Transplantation.

Minor histocompatibility antigens (mHAs), recipient-derived peptide epitopes presented on the cell surface, are known to mediate graft-versus-host disease (GVHD); however, there are no current methods to associate mHA features with GVHD risk. This deficiency is due in part to the lack of technological means to accurately predict, let alone confirm, the tremendous number of potential mHAs in each individual transplant. Previous studies have shown that different HLA molecules present varying fractions of candidate peptide epitopes; however, the genetic "distance" between HLA-matched donors and recipients is relatively constrained. From these 2 observations, it is possible that the HLA type for a donor-recipient pair (DRP) would provide a surrogate measurement of the number of predicted mHAs, which could be related to GVHD risk. Because different HLA molecules present variable numbers of peptide antigens, a predicted cumulative peptide-binding efficiency can be calculated for individual DRP based on the pair's HLA type. The purpose of this study was to test whether cumulative peptide-binding efficiency is associated with the risk of acute GVHD (aGVHD) or relapse. In this retrospective Center for International Blood and Marrow Transplant Research study, a total of 3242 HLA-matched DRPs were analyzed for predicted cumulative peptide-binding efficiency using their HLA types and were divided into tertiles based on their scores. Univariable and multivariable analyses was performed to test for associations between cumulative peptide-binding efficiency for DRPs, divided into the HLA-matched related donor (MRD) and HLA-matched unrelated donor (MUD) cohorts, and the primary outcomes of aGVHD and relapse. Secondary outcomes investigated included overall survival, disease-free survival, and transplantation-related mortality. Using a computationally generated peptidome as a test dataset, the tested series of HLA class I displayed peptide-binding frequencies ranging from 0.1% to 3.8% of the full peptidome, and HLA class II molecules had peptide-binding frequencies of 12% to 77% across the HLA-DRB1 allotypes. By increasing binding efficiency tertile, the cumulative incidence of aGVHD at 6 months for MUD patients was 41%, 41%, and 45% for HLA class I (P = .336) and 44%, 41%, and 42% for HLA class II (P = .452). The cumulative incidences of relapse at 3 years for MUD transplant recipients were 36%, 38%, and 38% for HLA class I (P = .533) and 37%, 37%, and 38% for HLA class II (P = .896). The findings were similar for MRD transplant recipients. Multivariable analysis did not identify any impact of peptide-binding efficiency on aGVHD or relapse in MUD or MRD transplant recipients. Whereas GVHD is mediated by minor antigen mismatches in the context of HLA-matched allo-HCT, peptide-binding efficiency, which was used as a surrogate measurement for predicted number of binding antigens, did not provide additional clinical information for GVHD risk assessment. The negative result may be due to the limitations of this surrogate marker, or it is possible that GVHD is driven by a subset of immunogenic mHAs. Further research should be directed at direct mHA epitope and immunogenicity prediction.

The discovery of the HLA-B*56:75 allele shows that HLA sequencing is important in all transplant settings.

HLA-B*56:75 has a nonsynonymous C to G substitution in codon 73 compared to HLA-B*56:01:01:02.

Optimal Donor for African Americans with Hematologic Malignancy: HLA-Haploidentical Relative or Umbilical Cord Blood Transplant.

Although hematopoietic cell transplantation from an HLA-matched unrelated donor is potentially curative for hematologic malignancies, survival is lower for African Americans compared with Caucasians. Because only approximately 20% of African Americans will have an HLA-matched unrelated donor, many of these patients undergo HLA-haploidentical relative or umbilical cord blood transplantation. In this study, we analyzed outcomes after HLA-haploidentical related donor (n = 249) and umbilical cord blood (n = 118) transplantations in African American patients with hematologic malignancy between 2008 and 2016. The predominant disease was acute myelogenous leukemia for recipients of both types of donor grafts. The incidences of grade II-IV and III-IV acute graft-versus-host disease were higher after umbilical cord blood transplantation compared with HLA-haploidentical relative transplantation (56% and 29%, respectively, versus 33% and 11%, respectively; P < .0001). The 2-year incidence of transplantation-related mortality adjusted for age and conditioning regimen intensity was higher after umbilical cord blood transplantation compared with HLA-haploidentical related donor transplantation (31% versus 18%; P = .008); however, there were no between-group differences in the 2-year adjusted incidence of relapse (30% versus 34%; P = .51), overall survival (54% versus 57%; P = .66), or disease-free survival (43% versus 47%; P = .46). Our findings show that the use of HLA-haploidentical and umbilical cord blood transplants expands the access to transplantation with comparable leukemia-free and overall survival for African Americans with hematologic malignancies.

The clinical impact of donor-specific antibodies on antibody-mediated rejection and long-term prognosis after heart transplantation.

PURPOSE OF REVIEW: Outcomes after cardiac transplantation have improved over past decades, but long-term graft survival remains limited in part because of uncertainty regarding clinical implications of donor-specific antibodies (DSAs). The purpose of this review is to consolidate recent advances in knowledge on the topic of DSA and their potential to impact long-term prognosis after heart transplantation. RECENT FINDINGS: The presence of persistent DSA increases the risk of poor outcome after heart transplantation, including development of antibody-mediated rejection (AMR), graft failure, cardiac allograft vasculopathy, and mortality. Importantly, different DSA vary in clinical significance. DSA capable of activating the complement cascade portend a higher risk of developing AMR. human leukocyte antigen class I and class II antigens are expressed differently within the heart, and so, clinical manifestations of class I and class II DSA vary accordingly. Further, compared with class I, class II DSA carry an increased risk of graft loss and mortality. When comparing preexisting DSA with formation of de-novo DSA, de-novo DSA are associated with worse outcome. SUMMARY: DSAs are generally associated worse long-term prognosis after heart transplantation but vary in their clinical significance. Recognition of specific risk profiles is essential for guiding posttransplant antibody management.

The safety and efficacy of clofarabine in combination with high-dose cytarabine and total body irradiation myeloablative conditioning and allogeneic stem cell transplantation in children, adolescents, and young adults (CAYA) with poor-risk acute leukemia.

Acute leukemias in children with CR3, refractory relapse, or induction failure (IF) have a poor prognosis. Clofarabine has single agent activity in relapsed leukemia and synergy with cytarabine. We sought to determine the safety and overall survival in a Phase I/II trial of conditioning with clofarabine (doses 40 - 52 mg/m2), cytarabine 1000 mg/m2, and 1200 cGy TBI followed by alloSCT in children, adolescents, and young adults with poor-risk leukemia. Thirty-seven patients; Age 12 years (1-22 years); ALL/AML: 34:3 (18 IF, 10 CR3, 13 refractory relapse); 15 related, 22 unrelated donors. Probabilities of neutrophil, platelet engraftment, acute GvHD, and chronic GvHD were 94%, 84%, 49%, and 30%, respectively. Probability of day 100 TRM was 8.1%. 2-year EFS (event free survival) and OS (overall survival) were 38.6% (CI95: 23-54%), and 41.3% (CI95: 25-57%). Multivariate analysis demonstrated overt disease at time of transplant (relative risk (RR) 3.65, CI95: 1.35-9.89, P = 0.011) and umbilical cord blood source (RR 2.17, CI95: 1.33-4.15, P = 0.019) to be predictors of worse EFS/OS. This novel myeloablative conditioning regimen followed by alloSCT is safe and well tolerated in CAYA with very poor-risk ALL or AML. Further investigation in CAYA with better risk ALL and AML undergoing alloSCT is warranted.

CPRA for allocation of kidneys in the US: More candidates ≥98% CPRA, lower positive crossmatch rates and improved transplant rates for sensitized patients.

In 2009 calculated panel reactive antibody (CPRA) replaced PRA as the metric for HLA sensitization in the US kidney allocation system. During the next four years, registrants with at least one unacceptable antigen increased (34-40%) and registrants with ≥98% PRA/CPRA increased from 7% to 9% of the waitlist. These changes were accompanied by a reduction in kidney offers refused for positive crossmatch: 14,137 (1.7%) in 2009 and 3,310 in 2013 (0.4%). Registrants with ≥98% PRA/CPRA had highest rates of refusal but also showed substantial improvement (20% in 2009 vs 8% in 2013). For registrants with ≥98% PRA/CPRA, 45% of accepted offers in 2009 were not transplanted into the intended recipient compared to 11% in 2013. Transplant rates remained low for these patients (∼50/1000 active patient-years), but rates improved for patients with 80-97% PRA/CPRA (223/1000 active patient-years in 2009 vs 354/1000 in 2013). In 2013, 40% regraft candidates had CPRA ≥98% compared to 4% of primary graft candidates. More females than males were ≥98% CPRA (14% vs 7%) and more females had CPRA above 0 (50% vs 28%). In the CPRA era, listing of unacceptable antigens increased, positive crossmatches were diminished and transplant rates for sensitized patients improved.

A Phase I Study of Reduced-Intensity Conditioning and Allogeneic Stem Cell Transplantation Followed by Dose Escalation of Targeted Consolidation Immunotherapy with Gemtuzumab Ozogamicin in Children and Adolescents with CD33+ Acute Myeloid Leukemia.

Myeloablative conditioning and allogeneic hematopoietic stem cell transplant (alloHSCT) in children with acute myeloid leukemia (AML) in first complete remission (CR1) may be associated with significant acute toxicity and late effects. Reduced-intensity conditioning (RIC) and alloHSCT in children is safe, feasible, and may be associated with less adverse effects. Gemtuzumab ozogamicin (GO) induces a response in 30% of patients with CD33+ relapsed/refractory AML. The dose of GO is significantly lower when combined with chemotherapy. We examined the feasibility and toxicity of RIC alloHSCT followed by GO targeted immunotherapy in children with CD33+ AML in CR1/CR2. Conditioning consisted of fludarabine 30 mg/m2 × 6 days, busulfan 3.2 to 4 mg/kg × 2 days ± rabbit antithymocyte globulin 2 mg/kg × 4 days followed by alloHSCT from matched related/unrelated donors. GO was administered ≥60 days after alloHSCT in 2 doses (8 weeks apart), following a dose-escalation design (4.5, 6, 7.5, and 9 mg/m2). Fourteen patients with average risk AML received RIC alloHSCT and post-GO consolidation: median age 13.5 years at transplant (range, 1 to 21), male-to-female 8:6, and disease status at alloHSCT 11 CR1 and 3 CR2. Eleven patients received alloHSCT from 5-6/6 HLA-matched family donors: 8 received peripheral blood stem cells, 2 received bone marrow, and 1 received related cord blood transplantation. Three patients received an unrelated allograft (two 4-5/6 and one 9/10) from unrelated cord blood unit and bone marrow, respectively. Neutrophil and platelet engraftment was observed in all assessable patients (100%), achieved at median 15.5 days (range, 7 to 31) and 21 days (range, 10 to 52), respectively. Three patients received GO at dose level 1 (4.5 mg/m2 per dose), 5 at dose level 2 (6 mg/m2 per dose), 3 at dose level 3 (7.5 mg/m2 per dose), and 3 at dose level 4 (9 mg/m2 per dose). Three of 14 patients received only 1 dose of GO after alloHSCT. One patient experienced grade III transaminitis, which resolved; no grade IV transaminitis, no grade III/IV hyperbilirubinemia, or sinusoidal obstructive syndrome were observed. The second dose of GO was given at median of 143 days (range, 120 to 209) after alloHSCT. Probability of grades II to IV acute and chronic graft-versus-host disease were 21% and 33.5%, respectively. Probability of overall survival after RIC alloHSCT and GO consolidation at 1 and 5 years was 78% and 61%, respectively. Probability of 5-year event-free survival after RIC alloHSCT and GO consolidation in patients in CR1 was 78%. No dose-limiting toxicities probably or directly related to GO were observed in this cohort. This preliminary data demonstrate that RIC followed by alloHSCT and consolidation with GO appears to be safe in children and adolescents with CD33+ AML in CR1/CR2. A phase II trial is currently underway investigating this approach with a GO dose of 9 mg/m2 per dose.

Pilot trial of risk-adapted cyclophosphamide intensity based conditioning and HLA matched sibling and unrelated cord blood stem cell transplantation in newly diagnosed pediatric and adolescent recipients with acquired severe aplastic anemia.

BACKGROUND: Cyclophosphamide-based conditioning regimens and allogeneic hematopoietic stem cell transplantation (AlloHSCT) from matched related donors (MRD) has resulted in the highest survival rates in children and adolescents with acquired severe aplastic anemia (SAA). Time to transplant has consistently been associated with decreased overall survival. Reduced toxicity conditioning and AlloHSCT has been used successfully in other pediatric non-malignant diseases. PROCEDURE: We piloted a risk-adapted AlloHSCT approach, using fludarabine and anti-thymocyte globulin based conditioning with high (200 mg/kg) and low (60 mg/kg) dose cyclophosphamide as upfront treatment in newly diagnosed pediatric patients with acquired SAA incorporating alternative donor sources, including cord blood. Average risk for non-engraftment patients with <10 transfusions received low dose cyclophosphamide (60 mg/kg); High Risk, those with ≥10 transfusions received conditioning regimen with higher intensity cyclophosphamide (200 mg/kg). RESULTS: Seventeen patients were enrolled and underwent AlloHSCT including 12 males and 5 females with mean age of 8 years (range 3-16), and median follow-up time of 39 months (range 1-135). Donor sources included MRD BM (6/6 [n = 9], 5/6 [n = 2]) and unrelated CB (5/6 [n = 4], 4/6 [n = 2]). Five year OS was 67.6% (37.9-85.4). Three secondary graft failures (17.6%) occurred in the low dose cyclophosphamide arm. CONCLUSIONS: Upfront treatment with risk-adapted cyclophosphamide conditioning AlloSCT is well tolerated for the management of newly diagnosed pediatric and adolescent patients with acquired SAA. However, the increased risk of graft rejection in the lower dose arm warrants additional research regarding the optimal intensity of cyclophosphamide-based conditioning regimen to reduce toxicity without increasing graft failure.

Identification of a permissible HLA mismatch in hematopoietic stem cell transplantation.

In subjects mismatched in the HLA alleles C*03:03/C*03:04 no allogeneic cytotoxic T-lymphocyte responses are detected in vitro. Hematopoietic stem cell transplantation (HSCT) with unrelated donors (UDs) showed no association between the HLA-C allele mismatches (CAMMs) and adverse outcomes; antigen mismatches at this and mismatches other HLA loci are deleterious. The absence of effect of the CAMM may have resulted from the predominance of the mismatch C*03:03/C*03:04. Patients with hematologic malignancies receiving UD HSCT matched in 8/8 and 7/8 HLA alleles were examined. Transplants mismatched in HLA-C antigens or mismatched in HLA-A, -B, or -DRB1 presented significant differences (P < .0001) in mortality (hazard ratio [HR] = 1.37, 1.30), disease-free survival (HR = 1.33, 1.27), treatment-related mortality (HR = 1.54, 1.54), and grade 3-4 acute graft-versus-host disease (HR = 1.49, 1.77) compared with the 8/8 group; transplants mismatched in other CAMMs had similar outcomes with HR ranging from 1.34 to 172 for these endpoints. The C*03:03/C*03:04 mismatched and the 8/8 matched groups had identical outcomes (HR ranging from 0.96-1.05). The previous finding that CAMMs do not associate with adverse outcomes is explained by the predominance (69%) of the mismatch C*03:03/03:04 in this group that is better tolerated than other HLA mismatches.

Impact of allele-level HLA matching on outcomes after myeloablative single unit umbilical cord blood transplantation for hematologic malignancy.

We studied the effect of allele-level matching at human leukocyte antigen (HLA)-A, -B, -C, and -DRB1 in 1568 single umbilical cord blood (UCB) transplantations for hematologic malignancy. The primary end point was nonrelapse mortality (NRM). Only 7% of units were allele matched at HLA-A, -B, -C, and -DRB1; 15% were mismatched at 1, 26% at 2, 30% at 3, 16% at 4, and 5% at 5 alleles. In a subset, allele-level HLA match was assigned using imputation; concordance between HLA-match assignment and outcome correlation was confirmed between the actual and imputed HLA-match groups. Compared with HLA-matched units, neutrophil recovery was lower with mismatches at 3, 4, or 5, but not 1 or 2 alleles. NRM was higher with units mismatched at 1, 2, 3, 4, or 5 alleles compared with HLA-matched units. The observed effects are independent of cell dose and patient age. These data support allele-level HLA matching in the selection of single UCB units.

Amino acid substitution at peptide-binding pockets of HLA class I molecules increases risk of severe acute GVHD and mortality.

HLA disparity has a negative impact on the outcomes of hematopoietic cell transplantation (HCT). We studied the independent impact of amino acid substitution (AAS) at peptide-binding positions 9, 99, 116, and 156, and killer immunoglobulin-like receptor binding position 77 of HLA-A, B, or C, on the risks for grade 3-4 acute graft-versus-host disease (GVHD), chronic GVHD, treatment-related mortality (TRM), relapse, and overall survival. In multivariate analysis, a mismatch at HLA-C position 116 was associated with increased risk for severe acute GVHD (hazard ratio [HR] = 1.45, 95% confidence interval [CI] = 1.15-1.82, P = .0016). Mismatch at HLA-C position 99 was associated with increased transplant-related mortality (HR = 1.37, 95% CI = 1.1-1.69, P = .0038). Mismatch at HLA-B position 9 was associated with increased chronic GVHD (HR = 2.28, 95% CI = 1.36-3.82, P = .0018). No AAS were significantly associated with outcome at HLA-A. Specific AAS pair combinations with a frequency >30 were tested for association with HCT outcomes. Cysteine to tyrosine substitution at position 99 of HLA-C was associated with increased TRM (HR = 1.78, 95% = CI 1.27-2.51, P = .0009). These results demonstrate that donor-recipient mismatch for certain peptide-binding residues of the HLA class I molecule is associated with increased risk for acute and chronic GVHD and death.