Table 2.

Multivariate analyses of patients with AML who underwent allogeneic HSCT

Variable	CIR		OS
	HR (95% CI)*	P	OR (95% CI)†	P
ELN2017 genetic-risk group (adverse vs intermediate vs favorable)	2.95 (2.09-4.18)	<0.001	0.69 (0.54-0.90)	.005
Age at HSCT, y	—	—	.95 (0.93-0.97)	<.001
Hemoglobin at diagnosis, g/dL	—	—	1.13 (1.0-1.26)	.04
Platelets at diagnosis, ×109/L	—	—	1.00 (1.00-1.01)	.06
Bone marrow blasts at diagnosis, %	—	—	1.02 (1.00-1.03)	.01
Peripheral blasts at diagnosis, %	—	—	0.99 (0.98-1.00)	.07

Variables considered in the models were those significant at α = 0.20 in univariate analyses. For the CIR end point, the variables were ELN2017 genetic-risk group (adverse vs intermediate vs favorable), hemoglobin at diagnosis, bone marrow blasts at diagnosis, age at HSCT, disease origin (secondary vs de novo), and remission status at HSCT (complete remission vs no complete remission). For the OS end point, the variables were ELN2017 genetic-risk group, hemoglobin at diagnosis, platelets at diagnosis, bone marrow blasts at diagnosis, peripheral blood blasts at diagnosis, age at HSCT, disease origin (secondary vs de novo), remission status at HSCT (complete remission vs no complete remission), HLA antigen match (antigen match vs mismatch), and application of the MAC regimen (MAC vs RIC vs NMA).

EVI1, ecotropic viral integration site-1; HR, hazard ratio.

^*Hazard ratio, <1 (>1) indicates a lower (higher) risk of relapse for the first category of dichotomous variable.

^†Odds ratio, <1 (>1) indicates a lower (higher) chance of survival for the first category of dichotomous variable.