Impact of donor KIR genotype and corresponding patients’ ligands on relapse incidence and mortality
| Classifier . | N . | % . | Relapse incidence . | Mortality . | ||
|---|---|---|---|---|---|---|
| HR (95% CI) . | P . | HR (95% CI) . | P . | |||
| KIR3DL1/HLA-B subtype combinations | ||||||
| Strong inhibiting KIR3DL1 | 574 | 26.0 | 1 | 1 | ||
| Weak inhibiting KIR3DL1 | 561 | 25.5 | 0.95 (0.75-1.21) | .7 | 0.87 (0.73-1.05) | .2 |
| Noninhibiting KIR3DL1 | 1069 | 48.5 | 1.11 (0.90-1.36) | .3 | 0.98 (0.84-1.14) | .7 |
| Strong inhibiting KIR3DL1 | 574 | 26.0 | 1 | 1 | ||
| Weak/noninhibiting KIR3DL1 | 1630 | 74.0 | 1.05 (0.87-1.28) | .6 | 0.94 (0.82-1.09) | .4 |
| KIR2DS1/C1C2 epitope combinations | ||||||
| KIR2DS1 absent | 1403 | 63.1 | 1 | 1 | ||
| KIR2DS1 present and C1+ | 699 | 31.5 | 0.97 (0.81-1.17) | .8 | 1.05 (0.91-1.21) | .5 |
| KIR2DS1 present and C2/C2 | 120 | 5.4 | 1.02 (0.70-1.48) | .9 | 1.13 (0.85-1.50) | .4 |
| KIR2DS1 absent and/or C2/C2 | 1523 | 68.5 | 1 | 1 | ||
| KIR2DS1 present and C1+ | 699 | 31.5 | 0.97 (0.81-1.17) | .8 | 1.04 (0.91-1.19) | .6 |
| Combined classifier based on donor KIR3DL1/HLA-B subtypes and KIR2DS1/C1C2 epitopes | ||||||
| Strong inhibiting KIR3DL1 | 432 | 19.6 | 1 | 1 | ||
| KIR2DS1 absent or C2/C2 | ||||||
| Strong inhibiting KIR3DL1 | 142 | 6.4 | 1.18 (0.80-1.75) | .4 | 1.47 (1.13-1.91) | .005 |
| KIR2DS1 present and C1+ | ||||||
| Weak/noninhibiting KIR3DL1 | 1081 | 49.0 | 1.13 (0.90-1.41) | .3 | 1.07 (0.90-1.27) | .5 |
| KIR2DS1 absent or C2/C2 | ||||||
| Weak/noninhibiting KIR3DL1 | 549 | 24.9 | 1.04 (0.80-1.34) | .8 | 0.99 (0.81-1.21) | .9 |
| KIR2DS1 present and C1+ | ||||||
| Impact of absence/presence of donor KIR3DS1 | ||||||
| KIR3DS1 absent | 1407 | 63.3 | 1 | 1 | ||
| KIR3DS1 present | 815 | 36.7 | 0.94 (0.79-1.12) | .5 | 1.03 (0.90-1.17) | .7 |
| KIR3DS1 absent | 1407 | 63.5 | 1 | 1 | ||
| One copy of KIR3DS1 | 709 | 32.0 | 0.91 (0.76-1.09) | .3 | 1.02 (0.88-1.17) | .8 |
| Two copies of KIR3DS1 | 100 | 4.5 | 1.18 (0.82-1.71) | .4 | 1.10 (0.81-1.48) | .5 |
| Classifier . | N . | % . | Relapse incidence . | Mortality . | ||
|---|---|---|---|---|---|---|
| HR (95% CI) . | P . | HR (95% CI) . | P . | |||
| KIR3DL1/HLA-B subtype combinations | ||||||
| Strong inhibiting KIR3DL1 | 574 | 26.0 | 1 | 1 | ||
| Weak inhibiting KIR3DL1 | 561 | 25.5 | 0.95 (0.75-1.21) | .7 | 0.87 (0.73-1.05) | .2 |
| Noninhibiting KIR3DL1 | 1069 | 48.5 | 1.11 (0.90-1.36) | .3 | 0.98 (0.84-1.14) | .7 |
| Strong inhibiting KIR3DL1 | 574 | 26.0 | 1 | 1 | ||
| Weak/noninhibiting KIR3DL1 | 1630 | 74.0 | 1.05 (0.87-1.28) | .6 | 0.94 (0.82-1.09) | .4 |
| KIR2DS1/C1C2 epitope combinations | ||||||
| KIR2DS1 absent | 1403 | 63.1 | 1 | 1 | ||
| KIR2DS1 present and C1+ | 699 | 31.5 | 0.97 (0.81-1.17) | .8 | 1.05 (0.91-1.21) | .5 |
| KIR2DS1 present and C2/C2 | 120 | 5.4 | 1.02 (0.70-1.48) | .9 | 1.13 (0.85-1.50) | .4 |
| KIR2DS1 absent and/or C2/C2 | 1523 | 68.5 | 1 | 1 | ||
| KIR2DS1 present and C1+ | 699 | 31.5 | 0.97 (0.81-1.17) | .8 | 1.04 (0.91-1.19) | .6 |
| Combined classifier based on donor KIR3DL1/HLA-B subtypes and KIR2DS1/C1C2 epitopes | ||||||
| Strong inhibiting KIR3DL1 | 432 | 19.6 | 1 | 1 | ||
| KIR2DS1 absent or C2/C2 | ||||||
| Strong inhibiting KIR3DL1 | 142 | 6.4 | 1.18 (0.80-1.75) | .4 | 1.47 (1.13-1.91) | .005 |
| KIR2DS1 present and C1+ | ||||||
| Weak/noninhibiting KIR3DL1 | 1081 | 49.0 | 1.13 (0.90-1.41) | .3 | 1.07 (0.90-1.27) | .5 |
| KIR2DS1 absent or C2/C2 | ||||||
| Weak/noninhibiting KIR3DL1 | 549 | 24.9 | 1.04 (0.80-1.34) | .8 | 0.99 (0.81-1.21) | .9 |
| KIR2DS1 present and C1+ | ||||||
| Impact of absence/presence of donor KIR3DS1 | ||||||
| KIR3DS1 absent | 1407 | 63.3 | 1 | 1 | ||
| KIR3DS1 present | 815 | 36.7 | 0.94 (0.79-1.12) | .5 | 1.03 (0.90-1.17) | .7 |
| KIR3DS1 absent | 1407 | 63.5 | 1 | 1 | ||
| One copy of KIR3DS1 | 709 | 32.0 | 0.91 (0.76-1.09) | .3 | 1.02 (0.88-1.17) | .8 |
| Two copies of KIR3DS1 | 100 | 4.5 | 1.18 (0.82-1.71) | .4 | 1.10 (0.81-1.48) | .5 |
Patients were classified according to the algorithms published by Boudreau et al19 in 2017, Venstrom et al18 in 2012, and Venstrom et al31 in 2010. Numbers of patients, grouped according to their donors’ KIR genotype, do not always add up to 100% because very few patients could not be classified according to the respective algorithm due to ambiguities. HRs were calculated in multivariable Cox regression models adjusted for patient age, donor age, disease risk index, Karnofsky performance status, sex match, CMV match, HLA match, conditioning intensity, T-cell depletion, and stem cell source.