Table 3.

Receiver operating characteristic (ROC) curve evaluation for all-cause mortality prediction

Prediction of all-cause mortality: ROC analysis	Area under the curve (AUC)	95 % CI	p
Age	0.642	0.604–0.678	0.003
NYHA class	0.690	0.653–0.725	<0.001
NT-pro- BNP	0.675	0.638–0.711	<0.001
fT3/fT4	0.684	0.647–0.719	<0.001
NT-pro-BNP & fT3/fT4 model	0.732	0.693–0.766	<0.001
Combined model	0.800	0.753–0.817	<0.001
ROC curves comparison – Combined model vs:	AUC (difference)	95 % CI	p
Age	0.150	0.064–0.235	<0.001
NYHA class	0.090	0.030–0.150	0.003
NT-pro-BNP	0.115	0.023–0.207	0.014
fT3/fT4	0.103	0.034–0.173	0.004
NT-pro-BNP & fT3/fT4 model	0.055	−0.032–0.142	0.218

CI confidence interval, NT-pro-BNP N-terminal pro-B-Type natriuretic peptide, fT3/fT4 Free Triiodothyronine/Free Thyroxine ratio, fT4 Free Thyroxine, NYHA New York Heart Association functional class; Combined model - composed by parameters of age, NYHA, NT-pro-BNP, and fT3/fT4