Abstract II: HLA-Centric Approaches in Graft Selection, Infectious Risk, and Immune Regulation
The Impact of HLA Mismatches on Survival after Unrelated Donor Hematopoietic Cell Transplantation Within Vivo T-Cell Depletion for Non-Malignant Disorders Varies by Patient Age and Diagnosis
Clinical Consulting Scientist IGEN/AFIP and Federal University of Parana Curitiba, Parana, Brazil
Aim: The role of in vivo T-cell depletion (TCD) in mitigating the impact of HLA mismatches in unrelated donor hematopoietic cell transplantation (HCT) for non-malignant disorders (NMD) remains unclear.
Methods: We used a public CIBMTR dataset (P-5679) to evaluate overall survival (OS) in 891 children, adolescents, and young adults with NMDs who received 8/8 matched unrelated donor (MUD) and mismatched (7/8 or 6/8) unrelated donor (MMUD) HCT with in vivo TCD from 2008 to 2017. Cox regression was used to assess OS. Covariates violating the proportionality assumption were adjusted via stratification. Interaction tests were performed between HLA matching and the other covariates. A random effect (frailty) was included in the Cox models to account for center effect. Statistical significance was defined as P ≤ 0.05.
Results: Donors were 659 MUD and 232 MMUD (94% 7/8). In vivo TCD consisted of ATG (n=502) or Campath (n=389). No interaction was found between HLA matching and TCD type (P=0.09), whereas significant interactions were observed with patient age (P=0.035) and severe aplastic anemia (SAA) diagnosis (P=0.006). Hence, Cox models were fit separately for adult (>18 years; n=212) and pediatric (≤18 years; n=679) cohorts and for SAA (n=310) and non-SAA (n=581) groups (Figure 1A). In multivariable Cox regression, stratified by graft source and adjusted for preparatory regimen, HCT-CI, year of HCT, and center effect, HLA mismatching was significantly associated with poorer OS in adults (HR=2.68; 95%CI=1.39-5.20; P=0.0034) but not in pediatric patients (HR=1.19; 95%CI= 0.79-1.78; P=0.40). Similarly, MMUD HCT was linked to worse OS in SAA patients (HR=3.00; 95%CI=1.59-5.64; P=0.0006), whereas no association was observed in non-SAA cohort (HR=1.04; 95%CI=0.68-1.59; P=0.85). Further analyses with four stratified groups (adult/SAA, n=153; pediatric/SAA, n=157; adult/non-SAA, n=59; pediatric/non-SAA, n=522) revealed a significant association between HLA mismatch and inferior OS (HR=4.11; 95%CI=1.73-9.74; P=0.0013) only in adults with SAA (Figure 1B).
Conclusion: Our results highlight that in vivo TCD minimizes the adverse effect of HLA mismatches in pediatric patients and adults without SAA diagnosis. Conversely, OS post-MMUD HCT for adults with SAA remains suboptimal, warranting further research to improve outcomes.