Table 4. Effect of Missing KIR Ligands on the Outcomes of Allogeneic HCT.
| Acute GVHD Grade II-IV | Chronic GVHD NST | Relapse | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Marker | N | p | SHR | 95%CI | p | SHR | 95%CI | p | SHR | 95%CI |
| One or more missing ligands for donor inhibitory KIR | ||||||||||
| Discovery cohort | 117/136 | 0.524 | 0.69 | 0.22–2.14 | 0.753 | 1.18 | 0.42–3.21 | 0.629 | 0.80 | 0.33–1.96 |
| Validation Cohort | 133/145 | 0.992 | 0.99 | 0.22–4.43 | 0.575 | 0.67 | 0.17–2.66 | 0.443 | 2.12 | 0.31–4.49 |
| All Cases | 250/281 | 0.725 | 0.85 | 0.35–2.05 | 0.740 | 1.14 | 0.52–2.49 | 0.955 | 0.98 | 0.45–2.14 |
| Donor Type | ||||||||||
| Matched Sibling Donors | 133/153 | 0.726 | 0.78 | 0.21–3.01 | 0.670 | 0.82 | 0.33–2.02 | 0.100 | 5.66 | 0.73–43.78 |
| Matched Unrelated Donors | 116/128 | 0.820 | 1.22 | 0.21–7.02 | 0.314 | 2.27 | 0.45–11.29 | 0.001 | 0.22 | 0.10–0.54 |
| One or more missing ligands for donor activating KIR | ||||||||||
| Discovery cohort | 39/91 | 0.763 | 0.87 | 0.35–2.15 | 0.498 | 0.75 | 0.33–1.68 | 0.444 | 0.74 | 0.34–1.61 |
| Validation Cohort | 41/86 | 0.736 | 1.161 | 0.48–2.82 | 0.439 | 0.73 | 0.32–1.61 | 0.276 | 1.69 | 0.66–4.34 |
| All Cases | 80/177 | 0.965 | 0.98 | 0.53–1.82 | 0.364 | 0.77 | 0.44–1.34 | 0.879 | 1.05 | 0.57–1.92 |
| Donor Type | ||||||||||
| Matched Sibling Donors | 45/91 | - | - | - | 0.673 | 0.83 | 0.35–1.96 | 0.917 | 0.95 | 0.42–2.17 |
| Matched Unrelated Donors | 35/86 | 0.688 | 1.21 | 0.47–3.08 | 0.654 | 0.79 | 0.30–2.12 | 0.726 | 1.18 | 0.46–3.01 |
Abbreviations: GVHD = Graft versus host disease; NST = needing systemic therapy; SHR = sub hazard ratio (sub-distributional hazard); N = number of observations (missing KIR ligands) out of number of recipients in the corresponding cohort; “-” denotes distributions in which statistical computation was not possible.
Missing ligands for inhibitory KIR included absence of at least one of A3/11, Bw4 (KIR3DL1 ligand), C1 (KIR2DL2/2DL3 ligand) or C2 (KIR2DL1 ligand) in recipients when the corresponding inhibitory receptors were present in their respective donors. Missing ligands for activating KIR included absence of either Bw4 (KIR3DS1 ligand) or C2 (KIR2DS1 ligand) in recipients when the corresponding activating receptors were present in the donor. Other poorly characterized activating KIR-ligand combinations, though dispersed in literature, were excluded from this analysis but were taken into consideration in supplementary data. A multivariate competing risks regression model was used to estimate the effect missing ligands on GVHD (acute and chronic) and relapse across discovery and validation cohorts individually and in combination (all cases). Significance of association was also tested separately among recipients of sibling and unrelated donor HCT. Sub-distributional hazard were described as sub-hazard ratios (SHR); p-values <0.05 were considered statistically significant (presented here in bold fonts).