Matched Sibling Donor Versus 8/8 Matched Unrelated Donor in Hematopoietic Cell Transplantation for Nonmalignant Disorders: A Propensity Score-Matched Analysis

Aim: In allogeneic hematopoietic cell transplantation (HCT), a matched sibling donor (MSD) is traditionally considered the first donor option, followed by 8/8 matched unrelated donors (MUD). However, in contemporary HCT for nonmalignant disorders, it is unclear whether a MUD yields similar overall survival (OS) compared to MSD.

Methods: To assess OS difference between MSD and MUD, we analyzed a public CIBMTR dataset (P-5679) with propensity score (PS) matching to account for confounding factors. Logistic regression was used to calculate PS. The PS matching applied a nearest-neighbor approach with a 1:1 matching ratio and a strict caliper value of 0.05. We utilized standardized mean difference (SMD) to assess covariate balance, with SMD <0.1 indicating good balance. Mann-Whitney test was used to compare median PS and SMD values. Kaplan-Meier curves, Log-rank test, and Cox regression were employed to evaluate OS following PS matching. Statistical significance was set at P<0.05.

Results: Overall, 1318 donor/recipient pairs were evaluated, with donor types comprising 598 MSD and 720 MUD. MSD and MUD were highly imbalanced before PS matching (median SMD=0.333, considering serotherapy, conditioning, graft source, GVHD prophylaxis, diagnosis, and patient race as confounders). After PS calculation, there were significantly different median PS values between donor types (MSD=0.468 versus MUD=0.614; P=5.72e-46; Figure 1A). The PS-matched cohort included 414 MSD and 414 MUD, which showed identical median PS values (0.477 for MSD and MUD; P=0.98, Figure 1B) and exhibited proper balance (median SMD=0.058 for the main confounders; Figure 1C). In the total PS-matched cohort, the 3-year OS probability was 87% (95% CI=84.4-89.3%). When stratified by donor type, the probability of 3-year OS was 90.6% (95% CI=87-93.2%) and 83.6% (95% CI=79.5-87%) for MSD and MUD, respectively (P=0.0054; Figure 1D). In Cox regression with PS matching, MUDs were associated with worse OS (HR=1.72; 95% CI=1.17-2.54; P=0.006).

Conclusion: MSD remains the gold standard in HCT for nonmalignant disorders due to its optimal OS. However, given the limited availability of MSDs, using a MUD is a viable option, as it also offers excellent OS post-HCT. Further studies are warranted to explore whether OS can be improved with innovative approaches, such as posttransplant cyclophosphamide.