A nomogram predictive model based on expressions of serum microRNA-30b-3p and soluble suppression of tumorigenicity 2 for poor prognosis in patients with severe heart failure
A nomogram predictive model based on expressions of serum microRNA-30b-3p and soluble suppression of tumorigenicity 2 for poor prognosis in patients with severe heart failure
Objective To analyze the expressions of serum microRNA-30b-3p (miR-30b-3p) and soluble suppression of tumorigenicity 2 (sST2),and to establish a nomogram predictive model for poor prognosis in patients with severe heart failure (HF).Methods The clinical materials of patients with severe HF (n=332,HF group) and healthy controls with physical examination (n=36,control group) were retrospectively analyzed in Department of Cardiovascular Medicine of Central Hospital of Suining City from June 2020 to June 2023.The AF group was divided,in a ratio of 7︰3,into modeling cohort (n=232) and model validation cohort (n=100).After followed up for 6 months,modeling cohort was divided,according to prognostic status,into modeling good prognosis group (n=162) and modeling poor prognosis group (n=70),and model validation cohort was divided into validation good prognosis group (n=70) and validation poor prognosis group (n=30).The differences in expressions of miR-30b-3p and sST2 and influence factors for poor prognosis were analyzed by using binary Logistic regression analysis.A predictive model was established and its efficacy was verified by using ROC curve.The consistency between predicted events and actual events was reviewed by using calibration curve.The effectiveness of the model was reviewed by using decision curve analysis (DCA).Results The levels of B-type natriuretic peptide (BNP),miR-30b-3p and sST2 were higher (t=18.829,t=11.282,t=8.313,all P=0.000),and left ventricular ejection fraction (LVEF) was lower (t=23.766,P=0.000) in HF group than those in control group.The levels of BNP,miR-30b-3p and sST2 were higher (t=11.734,t=14.181,t=13.218,all P=0.000),and LVEF was lower (t=13.285,P=0.000) in modeling poor prognosis group than those in modeling good prognosis group.The results of multi-factor Logistic analysis showed that increased BNP (OR=1.244,P=0.005),decreased LVEF (OR=0.432,P=0.001) and increased miR-30b-3p (OR=44.179,P=0.001) and sST2 (OR=10.125,P=0.003) were influence factors for poor prognosis.The predictive model equation based on results of multi-factor analysis was established as Log (P)=0.218×BNP-0.839×LVEF+3.788×miR-30b-3p+2.315×sST2-55.620.In modeling cohort AUC was 0.994 (95％CI:0.987～1.000),and in model validation cohort AUC was 0.964 (95％CI:0.924～1.000) in predicting survival rate by using nomogram.The difference between actual event curve and calibration curve in nomogram was small in modeling cohort and model validation cohort.The consistency between predicted events and actual events was higher (x2=1.206,P=0.997),and DCA also confirmed the effectiveness of the predictive mode to some extent.Conclusion The increased BNP,miR-30b-3p and sST2 are influence factors for poor prognosis in patients with severe heart failure.The nomogram predictive model for poor prognosis based on expressions of miR-30b-3p and sST2 shows good predictive ability,and consistency between predicted events and actual events is higher,which is worthy of clinical promotion and application.
关键词
心力衰竭/微小核糖核酸-30b-3p/可溶性致癌抑制因子2
Key words
Heart failure/MicroRNA-30b-3p/Soluble suppression of tumorigenicity 2
维普
万方数据