Allogeneic hematopoietic stem cell transplant (HSCT) is curative for patients (pts) with severe aplastic anemia (SAA). For SAA pts who lack HLA-identical donors, we explored a HSCT approach that co-infuses PBSCs enriched for CD34+ cells from a haplo-identical relative combined with a single umbilical cord blood unit (UCB). Although most pts undergoing this approach had early haplo-myeloid engraftment that was eventually supplanted by the UCB unit, engraftment patterns were highly variable; in some pts full cord myeloid chimerism was delayed greater than one year, while others had loss of cord engraftment with sustained full haplo-donor myeloid chimerism. Here, we investigated the impact of various patient, UCB, and haplo-donor characteristics, as well as NK cell KIR ligand mismatches between graft sources, on engraftment kinetics following haplo-cord HSCT.


Pts with SAA or SAA evolved to MDS unresponsive to immunosuppressive therapy with severe neutropenia (ANC<500), who lacked an HLA-matched donor and had an available haploidentical family member, and had at least one ≥ 4/6 matched UCB unit with a minimum TNC dose ≥1.5x107 cells/kg were eligible for HSCT. Pts were conditioned with cyclophosphamide (120 mg/kg), fludarabine (125 mg/m2), equine-ATG (160 mg/kg) and 200 cGy of total body irradiation. On Day 0, pts received a CD34 selected (Miltenyi CliniMacs) G-CSF mobilized haplo-donor allograft combined with a single UCB. Tacrolimus and MMF were given for GVHD prophylaxis.


16 pts (median age: 18.9 yrs, range: 4.5-27.9) including 13 SAA and 3 with SAA evolved to MDS underwent haplo-cord transplant. Ten pts received a 4/6, and six pts received a 5/6 HLA-matched UCB unit. A median 3.6x107 TNC/kg (range: 1.96-6.93) from the UCB unit, and 3.3x106CD34+ cells/kg (range: 3.0-4.1) from the haplo-donor were transplanted. All 16 pts had sustained engraftment with 15/15 pts evaluable past day 100 having transfusion independence. At a median follow-up of 570 days (range: 55-1826), 14/16 pts survive for an overall survival rate of 81.8%. Two pts died at day 414 and 402 post-HSCT from viral related complications (CMV pneumonitis and limbic encephalopathy). Neutrophil and platelet recovery occurred at a median 10 (range: 9-22), and 21 (range: 10-213) days post-HSCT, respectively. By post-HSCT day 11, 15/16 pts had neutrophil recovery. Cord myeloid engraftment (cord ANC>500, calculated from chimerism data) occurred in 13/16 pts at a median 42 days. 3/16 did not achieve a cord ANC>500 but had sustained haplo-donor engraftment. The cumulative incidence of acute grade II-IV GvHD was 38.1%. Engraftment profiles were highly variable among pts; 12 achieved full cord chimerism in all cell lineages, 2 remained mixed haplo-cord chimeras, and 2 failed to have UCB engraftment but had sustained 100% haplo-donor myeloid chimerism. Higher degrees of HLA matching (out of 10 alleles) between recipient and UCB unit were associated with faster rates of full cord engraftment (p=0.006) and a higher probability of complete loss of haplo-donor chimerism (p=0.018). KIR ligand incompatibility in the haplo vs. cord direction (defined as the presence of a KIR ligand in the haplo-donor graft that is absent in the UCB unit at HLA epitopes Bw4, HLA-C Group 1 & 2, HLA-A3, and HLA-A11) negatively impacted cord myeloid engraftment. 5/5 (100%) pts who failed to achieve full cord myeloid chimerism by post-HSCT day 400 had haplo vs. cord KIR ligand incompatibility. Moreover, both pts who failed to have UCB engraftment and had sustained haplo-donor chimerism had haplo vs. cord KIR ligand incompatibility. In contrast, only 3/11 (27%) pts who achieved full cord myeloid chimerism post-HSCT by day 231 had haplo vs. cord KIR ligand incompatibility (p=0.026). KIR ligand incompatibility in the cord vs. haplo direction showed no significant effect on haplo-donor myeloid engraftment.


These results show that haplo-cord HSCT is an effective treatment option for pts with SAA who lack an HLA-matched donor. Further, these results suggest that NK cell alloreactivity, occurring as a consequence of KIR ligand mismatch between the two graft sources, may have a negative impact on cord engraftment when haplo vs. cord KIR ligand incompatibility is present. In summary, this study highlights a novel factor that should be considered during graft selection for haplo-cord transplantation.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.