Clinical outcomes and prognostic factors in AML patients undergoing second allogeneic hematopoietic stem cell transplantation

Abstract Background/Aim A second allogeneic hematopoietic stem cell transplantation (SCT2) offers a potential cure for relapsed/refractory acute myeloid leukemia (AML), but is limited by suboptimal outcomes. With expanding donor options, evaluating outcome differences by donor types is essential. Previous studies have suggested that haploidentical transplantation (HIT) is associated with inferior outcomes compared to matched unrelated donors (MUD). In contrast, data comparing other donor types, such as matched sibling donors (MSD) and double cord blood (dCB), are limited. Methods We retrospectively reviewed 138 adult AML patients who underwent SCT2 at Catholic Hematology Hospital (Oct 2002-Feb 2025), extracting clinical data including demographics, disease status, hematopoietic stem cell transplantation-comorbity index (HCT-CI) scores, transplant history, and conditioning regimens. Conditioning intensity was categorized as high-, intermediate-, or low-intensity based on the Transplant Conditioning Intensity score: high-intensity regimens included Flu/Bu4, Bu-Cy, or TBI>1,200cGy; intermediate-intensity regimens included Flu/Bu2+TBI 800cGy; and low-intensity regimens included Flu/Bu2+TBI 400cGy. GVHD prophylaxis consisted of a calcineurin inhibitor and a short course of methotrexate. Anti-thymocyte globulin was administered in 95 patients (68.8%), while post-transplant cyclophosphamide was not used during the study period. GVHD prophylaxis included a calcineurin inhibitor and short-course methotrexate. Overall survival (OS) and relapse-free survival (RFS) were estimated via Kaplan-Meier analysis; relapse and non-relapse mortality (NRM) and acute and chronic GVHD (any grade) by cumulative incidence with competing risks. Group comparisons used log-rank or Gray's test. Prognostic factors for OS, RFS, relapse, and NRM were assessed by univariate and multivariate analyses; with variables (p<0.1) in univariate analysis were included in multivariate models. Results A total of 138 adult AML patients underwent SCT2 using MUD (n=55), MSD (n=14), HIT (n=41), or dCB(n=28). Median age at SCT2 was similar across groups (41.5–46 years), and most patients were in remission at SCT2 (85–100%). Baseline characteristics significantly differed by donor type, including SCT1 donor source (p<0.001) and conditioning intensity (p<0.001). High-intensity regimens were most frequently used in dCB recipients (92.9%), while intermediate-intensity regimens predominated in the HIT group (73.2%). MSD recipients had the highest proportion of patients with HCT-CI<3 (78.6%). Intensive salvage chemotherapy after SCT1 was more often used in MUD and dCB recipients (p=0.038). OS, RFS, NRM, and relapse incidence did not differ significantly by donor types. NRM tended to be higher in the dCB (60.7%), MSD (57.1%), and HIT (51.2%) groups compared to the MUD group (32.7%) (Gray's p=0.0914). Relapse incidence ranged from 21.4% (MSD, dCB) to 36.4% (MUD) (p=0.199), and RFS showed no significant group differences (p=0.695). Acute GVHD (any grade) occurred in 60.0% (MUD), 58.5% (HIT), 44.8% (dCB), and 35.7% (MSD) (p=0.286), while chronic GVHD (any grade) was most frequent in dCB (60.7%) and least in MSD (7.1%) (p=0.635). In multivariate analysis, longer SCT1–relapse interval (HR 0.988, 95% CI 0.978–0.999, p=0.036) and remission status at SCT2 (HR 3.485, 95% CI 1.922–6.320, p<0.001) were significantly associated with improved OS. For NRM, only SCT1–relapse interval was significant (HR 0.987, 95% CI 0.974–0.999, p=0.04). Relapse was associated with remission status at SCT2 (HR 5.873, 95% CI 2.621–13.160, p<0.001) and HCT-CI≥3 (HR 0.469, 95% CI 0.228–0.965, p=0.04). For RFS, both remission status at SCT2 (HR 5.873, 95% CI 2.621–13.160, p<0.001) and HCT-CI≥3 (HR 0.469, 95% CI 0.228–0.965, p=0.04) were significant. Conclusions Our findings revealed no statistically significant differences in survival outcomes of SCT2 according to donor type. However, disease status at SCT2 and the interval between SCT1 and relapse emerged as key prognostic factors. The overall suboptimal outcomes of allo-HCT2, driven by high rates of relapse and NRM, highlight the need for improved strategies, including optimized patient selection, enhanced GVHD prophylaxis, and post-transplant maintenance approaches.