. 2015 Jan 22;125(12):1922–1931. doi: 10.1182/blood-2014-08-596247

Table 2.

Prognostic significance of High-M and KM3 conditions

Variables	Kaplan-Meier log-rank test	Univariate Cox model	Multivariate Cox model (High-M, age 60 y, IILSS, 7q31.32 loss, 17p loss)	Multivariate Cox model (KM3, age 60 y, IILSS, 7q31.32 loss, 17p loss)
High-grade transformation^*	.0545^d	.0640^d
Increased LDH	.0426^a	.0323^a
7q31.32 loss	.0436^a	.0510^d	.1303	.1843
17p loss	.0713^d	.0259^a	.1731	.1314
Age <60 y	.0330^a	.0450^a	.0984	.1256
IILSS^†	.0133^b	.0398^a	.0122^a	.0148^a
High-M phenotype	.0032^b	.0048^b	.0143^a
KM3 phenotype	.0019^b	.0082^a		.0186^a
KLF4 (cg07309102)	.0031^b	.0046^b
CACNB2 (cg01805540)	.0022^b	.0034^b
HTRA1 (cg25920792)	.0030^b	.0045^b
Likelihood ratio test			.0001^c	.0002^c
Wald test			.0457^a	.0462^a
Log-rank test			.0005^b	.0007^b

LDH, lactate dehydrogenase.

Kaplan-Meier log-rank test and univariate Cox regression were performed for every variable. Multivariate Cox regression models were adjusted for either High-M and KM3 conditions or the significant variables in the univariate regression.

Few cases, statistical error for multivariate model.

^†

IILSS including hemoglobin, LDH, and albumin levels: a, P < .05, b, P < .005, c, P < .0005; d, borderline significance.