Abstract
Allogeneic hematopoietic cell transplantation (HCT) is a curative therapy for many Acute Myeloid Leukemia (AML) patients. There are not direct comparisons of results of HCT according to donor type in our region.
Objective: To evaluate the impact of donor type in different outcomes as overall survival (OS), non relapse mortality (NRM), relapse (Rel) and disease free survival (DFS) in HCT for AML patients. Secondary objective was to evaluate early variables like engraftment, Acute Graft versus Host Disease (aGVHD), CMV reactivations, fungal infections and bacteriemias.
Materials and Methods: A retrospective analysis of 250 medical records was conducted, addressing all patients who received an HCT for AML in seven Argentinian centers performing all three types of transplants, between January 2012 and December 2016. Mean patient´s age was 37 years; mean donors age 34 years; forty one percent (41%) of patients and 51% of donors were male; 89 patients (36%) received a matched related Allo-SCT (Sib), 78 (31%) an unrelated Allo-SCT (UD) and 83 (33%) a haploidentical Allo-SCT (Haplo). Pretrasplant HCT-CI score was low risk in 51% of the patients, 48% of the population were in CR 1 A myeloablative-conditioning (MAC) was used in 81% of the general population, peripheral blood stem cells were used in 85%.
OS and DFS were tested with Kaplan Meier and NRM and Rel with cumulative incidence. Dichotomic variables were tested with Chi2. We conducted a multivariate analysis (MVA) for time dependent variables with Cox Regression.
Results: Median follow up were 1.59 years. Patient and donor gender, age, HCT-CI and MAC were similar between the three type of donors., Haplo and UD groups were more likely to be in late stage compared to Sib (64% and 62% vs. 36%; p<0.001). BM source was more frequently used in Haplo (27% vs. 8% Sib and 6% UD ; p=0.001). Engraftment was slower in the Haplo group than in Sib and UD (≥15 days) (85% vs. 35% vs. 40% respectively; p=0.001). Graft failure rates were higher in Haplo than Sib and UD (16% vs. 1% and 4%; p=0.001). CMV reactivation was also more frequent in the Haplo and UD setting (53% and 56% vs.34%; p=0.02). There was a trend to a higher aGVHD incidence in UD group (60% vs. 43% Sib vs 42% Haplo; p=0.06), with no differences in GII-IV (41% vs. 27% vs 33%; p=0.19) or cGVHD (27% vs. 22% vs. 25%; p=0.65).
Early NRM (day 100) was lower in Sib (4% Sib vs 13% UD vs 12% Haplo) as well as NRM at 1-3 years, (12-15% vs 21-25% vs 27-29% respectively; p=0.04) (Figure 1). Relapse rates (1-3 years) were higher among Sib and Haplo (39-44% Sib vs 40-40% Haplo vs 21-29% UD; p=0.05) (Figure 2). DFS (1-3 years) was higher in Sib and UD compared to Haplo (50-42% vs 57-46% vs 32-31% respectively; p=0.03) (Figure 3) as was OS (1-3 years 61-51% Sib vs 65-51% UD vs 44-34% Haplo; p=0.01) (Figure 4). This difference remained significant in the early stage disease (p=0.04) but no for late stage
Multivariate analysis (expressed as Hazard Ratios with 95% confidence interval) showed that pretrasplant intermediate or high risk HCT-CI score were associated with poorer OS (HR 2.40, 1.34 - 4.30) and (2.67, 1.20 - 5.95). No association was seen with the type of donor used. Late stage disease at diagnosis was associated with higher relapse rates, independent of the type of donor and type of conditioning regime used (HR 2.05, 1.16-3.64). No factor affected negatively NRM after adjustment in MVA.
Conclusion: Although early toxicity was similar in both Haplo and UD when expressed as NRM (day 100), worse 1 y and 3 y NRM and higher rates of relapse were observed in Haplo, resulting in a negative impact in OS. This experience shows that Haplo HCT is a complex procedure in the real world. Larger number of patients and prospective trials are needed to establish firm conclusions regarding the best strategy for patients with AML requiring a HCT
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.