Myocardial fibrosis combined with NT-proBNP improves the accuracy of survival prediction in ADHF patients

Table 5 Predictive improvement in the models for all-cause mortality

Model	NRI	95% CI	P	IDI	95% CI	P
Model 2	0.68*	0.36–1.01	< 0.01	0.12*	0.056–0.18	< 0.01
Model 3	0.64*	0.31–0.97	< 0.01	0.14*	0.072–0.21	< 0.01
Model 3	0.10**	− 0.25 to 0.46	0.56	0.021**	− 0.0042 to 0.046	0.10
Model 4	0.52*	0.20–0.84	< 0.01	0.17*	0.10–0.24	< 0.01
	0.31**	0.053–0.56	0.018	0.068**	0.024–0.11	< 0.01
	0.20***	− 0.12 to 0.52	0.23	0.054***	0.0065–0.10	0.026
Model 5	0.52*	0.20–0.84	< 0.01	0.17*	0.10–0.24	< 0.01
	0.31**	0.053–0.56	0.018	0.068**	0.024–0.11	< 0.01
	0.20***	− 0.12 to 0.52	0.23	0.054***	0.0068–0.10	0.025
	0.011****	− 0.36 to 0.40	0.95	0.00	–	1.0

Bold indicates the siginificant values (P < 0.05)
Model 1: Clinical factors contained age, gender, NYHA class, AF, SCr and LVEF; Model 2: Model 1 + Log2 NT-proBNP; Model 3: Model 2 + Log2 sST2; Model 4: Model 2 + Log2 PIIINP; Model 5: Model 2 + Log2 sST2 + Log2 PIIINP
CI, confidence interval; NRI, net reclassification improvement; IDI, integrated discrimination improvement
^*Compared with Model 1; **compared with Model 2;***compared with Model 3; ****compared with Model 4

ISSN: 1471-2261