Better Outcome Following Younger Haploidentical Donor Versus Older Matched Unrelated Donor Transplant for Fit Patients With Acute Myeloid Leukemia Transplanted in First Remission: A Study From the Global Committee and the Acute Leukemia Working Party of the EBMT

Abstract

Donor selection has a major impact on the results of allo-HCT. Human leukocyte antigen (HLA)-matched sibling donors (MSD) are considered preferred donors, but they are available for only a minority of patients. 10/10 HLA fully-matched unrelated donors (MUDs) or haploidentical donors (HAPLOs) are both important alternatives in the absence of MSDs, with growing evidence indicating comparable transplant outcomes between these two [1]. Older donor age has been associated with inferior transplant outcomes in either MUD [2] or HAPLO [3] settings essentially due to a higher rate of non-relapse mortality (NRM). In a recent EBMT study focusing on donor selection in the unrelated donor setting, older donor age but not HLA mismatch was associated with a higher relapse incidence (RI) and poorer survival outcomes [4].

In order to compare the transplant outcomes of younger HAPLO and older MUD allo-HCT and provide rational suggestions for donor selection, we compared in an EBMT global multi-center registry-based analysis the outcomes following allo-HCT from younger HAPLO or older MUD.

Data of 3296 adult AML patients receiving a first allo-HCT in CR1 from between 2010 and 2022 were analyzed. Transplants from mismatched UD (< 10/10) were excluded. The young HAPLO group was identified as transplants with haploidentical donors aged between 15 and 40 years old. The older MUD group was identified as transplants with 10/10 HLA fully-MUDs aged between 40 and 65 years old.

Overall survival (OS) was defined as the duration from transplant to death from any cause. Leukemia-free survival (LFS) was defined as the time to either relapse or death. GVHD-free-relapse-free survival (GRFS) was measured as the duration from transplantation to the earliest occurrence of one of the following events: grade III–IV aGVHD, extensive cGVHD, relapse, or death due to any cause. The diagnosis and grading of acute and chronic GVHD were conducted according to the revised criteria of the Mount Sinai International Consortium and the National Institutes of Health (NIH), respectively. Neutrophil recovery was characterized by maintaining an absolute neutrophil count above 0.5 × 10⁹/L for at least three consecutive days, while platelet recovery required maintaining a platelet count above 20 × 10⁹/L for the same duration, without the need for transfusion. To compare patient characteristics between groups, quantitative variables were analyzed using the Mann–Whitney U test, while categorical variables were assessed through chi-squared or Fisher's exact tests. The Kaplan–Meier estimator was utilized to calculate OS, LFS, and GRFS. Cumulative incidence calculations were used to determine RI, NRM, and the recovery rates of neutrophils and platelets, as well as GVHD outcomes. RI and NRM were treated as mutually competing events. Relapse and death were considered competing events for GVHD outcomes, while death was considered a competing event for neutrophil and platelet recovery.

Statistical analysis included univariable and multivariable analysis (MVA) adjusted for potential confounding factors using a Cox proportional-hazards regression model for main outcomes. Multivariable results were presented as hazard ratios (HR) with 95% confidence intervals (95% CI). Considering the missing data, all patients with data missing for the variable were not included in the MVA. Survival rates, cumulative incidence figures, and HR were presented along with their respective 95% confidence intervals (95% CI). All statistical analyses were performed using a two-sided approach, with the significance level established at 0.05. Statistical analyses were performed using GraphPad Prism version 9.0, SPSS version 26.0, and R 4.3.2.

Data of 2127 patients receiving younger HAPLO (15–40 years old(yo)) and 1169 patients receiving older MUD (40–65 yo) HCTs were analyzed. The median age of the overall patient cohort was 56.8 years, ranging from 18 to 75 years. Distributions of the disease risk classification ELN 2017 and measurable residual disease prior to transplantation were similar between the two groups. A greater proportion of patients in the younger HAPLO cohort underwent myeloablative conditioning (60%) compared to those in the older MUD cohort (51%, p < 0.001). Post-transplant cyclophosphamide was predominantly used as GVHD prophylaxis in the younger HAPLO group (86%), whereas ATG constituted the primary prophylactic regimen in the older MUD group (90%) (Table S1). The study was conducted as per the Declaration of Helsinki and Good Clinical Practice guidelines.

The median duration of follow-up was 2.2 years (IQR = 2.1–2.4) for the younger HAPLO group and 2.7 years (IQR = 2.4–3) for the older MUD group. Due to the difference of the follow-up between group, the follow-up were censured at 2 years. The estimated 2 year OS was 66.9% (95% CI, 64.5–69.1) for younger HAPLO as compared to 63.8% (95% CI, 60.6–66.8) for older MUD. The estimated 2 year LFS was 62.3% (95% CI, 59.9–64.7) for younger HAPLO and 57% (95% CI, 53.6–60.2) for older MUD. Finally, the 2 year cumulative incidence of relapse was 20% (95% CI, 18.1–22) for younger HAPLO versus 24.9% (95% CI, 22.1–27.8) for older MUD (Table S2).

We found that the proportion of patients with a Karnofsky ≥ 90 was higher for older MUD. To avoid any bias due to this interaction between variables, we separated the analysis according to the Karnofsky score (≥ 90 and < 90) during multivariate analysis (Table S3). The major transplant outcomes and complications are illustrated in Figures 1 and S1.

FIGURE 1

Open in figure viewerPowerPoint

GRFS (A), OS (B), LFS (C), RI (D), and NRM(E) comparing younger haploidentical (Haplo) and older matched unrelated donor (MUD) for AML patients allografted with Karnofsky score of either ≥ 90 or < 90.

MVA revealed that, for patients with a Karnofsky score ≥ 90, older MUD was linked to an increased risk of relapse (HR = 1.36, 95% CI 1.1–1.69; p = 0.004), cGVHD (HR = 1.28, 95% CI 1.07–1.55; p = 0.008) and extensive cGVHD incidence (HR = 1.83, 95% CI 1.35–2.48; p < 0.001), ultimately resulting in poorer LFS (HR = 1.2, 95% CI 1.03–1.41; p = 0.024) outcomes relative to younger HAPLO.

For patients with Karnofsky score < 90, although older MUD was associated with higher overall cGVHD (HR = 1.48, 95% CI 1.05–2.07; p = 0.024), it was associated with slightly better OS (HR = 0.75, 95% CI 0.57–1; p = 0.046) probably due to a trend toward better NRM (HR = 0.71, 95% CI 0.49–1.02; p = 0.065). There was no significant difference between the two groups concerning RI, aGVHD incidences, and LFS.

Additionally, Older patient age (per 10 years) was associated with higher NRM for both patients with Kanorfsfy score < 90 and ≥ 90, which led to inferior LFS and OS for those with Kanorfsfy score ≥ 90. Adverse ELN2022 cytogenetics was associated with higher RI, poorer LFS, and OS. For patients with Kanorfsfy score ≥ 90, secondary AML correlated with higher RI and worse OS compared with de novo AML.

We analyzed 3296 adult AML patients in the current study and observed better LFS following younger HAPLO versus older MUD transplant for adult AML patients with Karnofsky ≥ 90 transplanted in CR1.

In the current study, the LFS advantage of younger HAPLO over older MUD, as shown in the multivariate analysis, is primarily due to reduced RI. There has been evidence showing that HAPLO might exert superior graft-versus-leukemia effects compared to MUD transplants, especially in high-risk AML. Our two recent studies comparing HAPLO and MUD for either high-risk KMT2A rearranged AML [5] or CBF AML transplanted in CR2 [6] both showed significantly lower RI for HAPLO than for MUD, further supporting the advantage of HAPLO over MUD in disease relapse prevention.

The adverse effects of older donors on transplant outcomes were majorly documented as associated with a higher rate of NRM, while in the current study, younger donor age contributed essentially to better relapse prevention. The protective effect of younger donors against relapse has already been described in acute myeloid leukemia [4] and myelodysplastic syndrome [7]. The mechanisms by which donor age might affect the risk of relapse remain to be elucidated. The delay in immune recovery observed with older donors, along with the aging of T cells resulting in diminished graft-versus-leukemia efficacy [4], could be contributing factors.

For patients with Karnofsky score < 90, younger HAPLO was associated with slightly lower OS and higher NRM compared with older MUD, which might be related to lower tolerability of these patients to MAC conditioning, which is more frequently used in younger HAPLO allo-HCTs. Furthermore, the advantage of relapse prevention for younger HAPLO was not significant for patients with Karnofsky score < 90, which might also be due to the decreased use of myeloablative conditioning in this entire subgroup.

Regardless of Karnofsky score, older MUD was associated with higher incidences of cGVHD than younger HAPLO in the entire cohort. Donor type, older donor age, and minor HLA mismatch may have all contributed to this disparity. A previous EBMT study comparing HAPLO and MUD in elderly AML patients observed a higher risk of extensive cGVHD for MUD without differences for other outcomes [8]. Grunwald et al. in a CIBMTR study also found that chronic GVHD occurred more often in MUD than HAPLO without affecting survival outcomes [9]. Furthermore, higher incidences of both acute and chronic GVHD [2] with increasing donor age have been observed in a number of previous studies. A recent study based on the CIBMTR registry comparing younger HAPLO (≤ 35 years) and older MUD (> 35 years) allo-HCT in AML and MDS patients observed superior OS for younger HAPLO.

Our present study, with a big cohort of AML patients and detailed subgroup analysis based on Karnofsky score, strongly advocates for the selection of younger HAPLO donors over older MUD for fit AML patients demonstrating a Karnofsky score of 90 or higher.

Of course, there are limitations in our study that are inherent to its retrospective and registry-based nature. While the conditioning regimen for allo-HCT and the GVHD prophylaxis were heterogeneous, it is however noteworthy that the conditioning intensity (MAC vs. RIC) did not significantly affect any of the major transplant outcomes in either of the two Karnofsky score groups.