Table 2 Multivariate analysis of covariates associated with therapy failure (Fine-Gray regression).

Model 1 (reported in Zhang et al., training cohort)
Covariate	Regression coefficient	Hazard ratio (95% CI)	p-value
Male sex (binary)	0.1919	1.2 (1.0–1.5)	0.01
Age/100, years (continuous)	1.6160	5.0 (2.5–10.3)	<0.001
(Hemoglobin/100)^–2, g/L (continuous)	0.3105	1.4 (1.2–1.5)	<0.001
Blood blasts, % (continuous)	0.1087	1.1 (1.1–1.2)	<0.001
Spleen size below the costal margin, cm(continuous)	0.0671	1.1 (1.0–1.1)	<0.001
High-risk ACAs (binary)	0.5461	1.7 (1.1–2.8)	0.02
Model 2 (French cohort)
Covariate	Regression coefficient	Hazard ratio (95% CI)	p-value
Male sex (binary)	0.5042	1.66 (1.18–2.32)	0.0032
Age/100, years (continuous)	0.3609	1.43 (0.442–4.66)	0.55
(Hemoglobin/100)⁻², g/L (continuous)	0.0450	1.05 (0.629–1.74)	0.86
Blood blasts, % (continuous)	0.0381	1.04 (0.968–1.11)	0.29
Spleen size below the costal margin, cm (continuous)	0.0107	1.01 (0.972–1.05)	0.59
High-risk ACAs (binary)	0.4360	1.55 (0.933–2.56)	0.091
Model 3 (French cohort)
Covariate	Regression coefficient	Hazard ratio (95% CI)	p-value
Male sex (binary)	0.4817	1.62 (1.16–2.26)	0.0045
High-risk ACAs (binary)	0.4252	1.53 (0.947–2.47)	0.082
White blood cells, ×10E + 9/L (continuous)	0.0015	1.00 (1.00–1.00)	0.034
Comorbidities (binary)	0.2609	1.32 (0.968–1.74)	0.081

Model 1 considers the six variables selected by Zhang et al. in the Fine-Gray regression. The indicated values are the ones reported in the original study (their Table 2). Model 2 considers the same six variables, but the Fine-Gray regression coefficients were recalculated using data from our French cohort. Model 3 is the result of variable selection on a full Fine-Gray regression considering the twelve candidate variables in Zhang et al., computed using data from the French cohort.
ACAs additional chromosomal abnormalities.

Search