Outcomes of Graft Failures after Umbilical Cord Blood Transplantation in Acute Leukemia: A Study from Eurocord and the Acute Leukemia Working Party of the EBMT

Backround & Methods. Approximately 12% of adult patients (pts) receiving umbilical cord blood transplantation (UCBT) as treatment for acute myeloid (AML) or lymphoblastic (ALL) leukemia are reported to experience graft failure (Baron et al., J Hematol Oncol 2015 ; 8 :107). No systematic large study has assessed their outcome thus far. Therefore the acute leukemia working party of the EBMT and EUROCORD performed a multicenter retrospective registry study on adult patients with AML or ALL receiving a first single or double UCBT between 2005 and 2016.

Results. The study included 1836 pts. 215 of them (11.7%) had graft failure; 160 with AML and 55 with ALL. Their median age was 43 (range, 18-68) years. 115 pts were male and 100 female. Median year of transplantation was 2010. 67% of the patients were in complete remission (CR) at transplantation, while 33% had advanced disease. Conditioning regimen was myeloablative in 128 pts (60%) and RIC in 87(40%) pts, respectively. ATG was administered in 101 pts. 126 pts received a single UCBT and 49 a double UCBT. Among the 215 pts with graft failure 54 underwent a second allogeneic (n=53) or autologous (n=1) transplantation 16 to 700 (median 62) days after the first UCBT (the 1-year cumulative incidence of second transplantation was 24%). 100-day and 1 year OS in pts not undergoing a second transplantation were 21.9 % (95% CI: 16.3-29.3) and 10.6 % (95% CI: 6.7-16.8), respectively. Among the 53 pts who underwent a second allogeneic transplantation, the second donor was either an unrelated donor (n=34), an HLA-haploidentical (n=17), or an HLA-identical sibling (n=2). Stem cell source consisted of single (n=13) or double (n=13) UCB, PBSC (n=17), BM (n=7), PBSC + UCB (n=3) or PBSC + BM (n=1). Conditioning regimen for the second transplantation was a RIC in 87% of pts. Sustained engraftment was achieved by 33 pts (61%) after the second transplantation. One- and 3-year incidences of relapse, nonrelapse mortality, OS and LFS from second transplantation were 21% and 28%, 47% and 52%, 40% and 23%, and 33% and 21%, respectively. Main causes of death following second transplantation (n=39) included infection (n=14) and leukemia relapse/progression (n=12). In the 17 patients receiving an haploidentical transplantation as second graft 1- and 3-year OS were 51.8 % (95% CI: 32.4-82.7) and 41.4 % (95% CI: 21.8-78.6), respectively. In a multivariate Cox model assessing factors predicting OS in UCBT patients with graft failure, transplantation for advanced disease (HR=2.3; 95% CI : 1.6-3.3, P<0.0001) as well as graft failure after the salvage transplantation (HR=2.8; 95% CI : 1.6-5.0, P=0.0005) were predictive of poor OS while use of RIC conditioning for the first transplantation was associated with a better OS.

Conclusions. In this large cohort of adult pts given UCBT as treatment for AML or ALL the incidence graft failure was 12%. A second transplantation could be performed only to a minority (24%) of patients with high, about 50%, TRM but led to a 3-year OS of 23%, rescuing about quater of the pts with graft failure post UCBT. Results were more encouraging in the 17 patients who received a salvage HLA-haploidentical transplantation (3-year OS of 41%). Future attempts should be focused on reducing the high TRM associated with the second transplant while recent progresses in cord blood engineering will likely dramatically decrease the incidence of graft rejction after UCBT.