Umbilical cord blood (UCB) is frequently considered as a suitable alternate source of hematopoietic stem cells (HSC) for both pediatric and adult patients who require HSC transplant for treatment of high-risk or relapsed hematologic malignancy. As mismatched killer Ig-like receptor ligands (KIR-L MM) has been associated with anti-host alloreactive natural killer (NK) cell activity and reduced risk of acute myeloid leukemia relapse in recipients of T-cell depleted haploidentical HSC, we hypothesized that GVL after UCBT may also be mediated by NK alloreactivity. We therefore assessed the effect of KIR-L MM in 243 recipients of UCB transplanted at the University of Minnesota between 1998 and 2004 for whom HLA-A, -B, -C and DRB1 typing was available for both patient and UCB unit(s). Median age, weight, and follow-up were 27 yrs (range, 0.2–69), 64.6 kg (range, 3.8–120.2), and 1.1 yr (range, 0.5–6.6), respectively. For recipients of double UCBT (n =106), we analyzed the KIR-L assignment of the engrafting unit only. KIR-L MM in the GVH direction was established using the algorithm of Ruggeri and Velardi and was found in 70 (29%) donor-recipient pairs. Probability of 2-year survival was 54% (95%CI: 42–66) vs. 47% (95%CI: 38–56) (p=0.88) in KIR-L MM vs. matched pairs, respectively. Intensity of the conditioning regimen was the only independent predictor of survival, with a RR of death 1.54 (95%CI: 1.07–2.22, p=0.02) for those who received a non-myeloablative conditioning. For the whole group, incidence of relapse at 1-yr was 17% (95%CI: 8–26) vs. 31% (95%CI: 24–38) (p=0.12); for myeloid malignancy patients, incidence of relapse at 1 year was 13% (95%CI:2–24) vs. 34% (95%CI:22–46)(p=0.10). Intensity of the conditioning regimen was the only independent predictor of relapse, with a RR 2.09 (95%CI: 1.31–3.35, p<0.01) for those who received a non-myeloablative conditioning. Incidence of graft failure and grade II–IV acute GVHD was 11% (95%CI: 3–18) vs. 10% (95%CI: 6–15) (p= 0.97) and 50% (95%CI: 37–63) vs. 50% (95%CI: 42–58) (p=0.67), respectively. Notably, the incidence of TRM was higher among recipients of KIR-L MM grafts [26% (95%CI: 15–26) vs. 13% (95%CI: 8–18), p=0.01]. KIR-L MM did not impact on outcome within the subgroups of patients based on conditioning, number UCB units (single vs. double), and those who had a myeloid malignancy (AML, CML or MDS). In multivariate analysis, a nucleated cell dose < 3.5 X 10E7/kg [RR 3.86 (95%CI: 1.84–8.12, p<0.01)] and KIR-L MM [RR 2.15 (95%CI: 1.15–4.01, p<0.01)] were independent predictors of increased TRM. In summary, in the setting of UCBT, KIR-L MM is associated with increased TRM with no obvious beneficial effect on engraftment, relapse risk or survival as previously demonstrated in recipients of TCD haploidentical HSC. Differences in NK cells reconstitution, presence of T-cells and/or use of immunosuppression may interfere with any potential beneficial effect alloreactive NK cells in the setting of UCBT. While additional studies are still needed, results to date fail to support a specific search for UCB units with a KIR-L MM.

Author notes

Corresponding author