Table 2. Multivariable regression for... | American Society of Hematology

Table 2.

Multivariable regression for relapse in 3 candidate KIR genotypes

Model	G5		G3/HLA-Bw4		G2/HLA-C1C1
Model	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P
CIBMTR	G5, n = 136 vs all others, n = 1881		G3, n = 53 vs other Bw4, n = 1143		G2, n = 86 vs other C1C1 n = 761
Base model∗	0.52 (0.35, 0.78)	.002	0.32 (0.14, 0.72)	.006	1.63 (1.15, 2.29)	.005
Int∗ T-cell depletion†	0.55 (0.34, 0.92)	.021	0.50 (0.21, 1.21)	.13	1.35 (0.81, 2.27)	.30
Int∗ cond intensity	0.43 (0.23, 0.78)	.005	0.17 (0.04, 0.68)	.013	1.93 (1.17, 3.18)	.010
DRST	G5, n = 73 vs all others, n = 743		G3, n = 24 vs other Bw4, n = 451		G2, n = 42 vs other C1C1, n = 264
	1.01 (0.55, 1.85)	.97	0.87 (0.35, 2.17)	.76	0.41 (0.18, 0.96)	.041

Model	G5		G3/HLA-Bw4		G2/HLA-C1C1
Model	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P
CIBMTR	G5, n = 136 vs all others, n = 1881		G3, n = 53 vs other Bw4, n = 1143		G2, n = 86 vs other C1C1 n = 761
Base model∗	0.52 (0.35, 0.78)	.002	0.32 (0.14, 0.72)	.006	1.63 (1.15, 2.29)	.005
Int∗ T-cell depletion†	0.55 (0.34, 0.92)	.021	0.50 (0.21, 1.21)	.13	1.35 (0.81, 2.27)	.30
Int∗ cond intensity	0.43 (0.23, 0.78)	.005	0.17 (0.04, 0.68)	.013	1.93 (1.17, 3.18)	.010
DRST	G5, n = 73 vs all others, n = 743		G3, n = 24 vs other Bw4, n = 451		G2, n = 42 vs other C1C1, n = 264
	1.01 (0.55, 1.85)	.97	0.87 (0.35, 2.17)	.76	0.41 (0.18, 0.96)	.041

Cond, conditioning; int, interaction between donor KIR and recipient-HLA.

The base model for relapse includes the genotype and age, as selected by backward variable selection (see “Methods”).

Interactions with T-cell depletion and conditioning intensity were nonsignificant in all analyses.