Construction of a web‐based dynamic nomogram for predicting the prognosis in acute heart failure

Abstract

AbstractAimsThe early identification and appropriate management may provide clinically meaningful and substained benefits in patients with acute heart failure (AHF). This study aimed to develop an integrative nomogram with myocardial perfusion imaging (MPI) for predicting the risk of all‐cause mortality in AHF patients.Methods and resultsProspective study of 147 patients with AHF who received gated MPI (59.0 [47.5, 68.0] years; 78.2% males) were enrolled and followed for the primary endpoint of all‐cause mortality. We analysed the demographic information, laboratory tests, electrocardiogram, and transthoracic echocardiogram by the least absolute shrinkage and selection operator (LASSO) regression for selection of key features. A multivariate stepwise Cox analysis was performed to identify independent risk factors and construct a nomogram. The predictive values of the constructed model were compared by Kaplan–Meier curve, area under the curves (AUCs), calibration plots, continuous net reclassification improvement, integrated discrimination improvement, and decision curve analysis. The 1, 3, and 5 year cumulative rates of death were 10%, 22%, and 29%, respectively. Diastolic blood pressure [hazard ratio (HR) 0.96, 95% confidence interval (CI) 0.93–0.99; P = 0.017], valvular heart disease (HR 3.05, 95% CI 1.36–6.83; P = 0.007), cardiac resynchronization therapy (HR 0.37, 95% CI 0.17–0.82; P = 0.014), N‐terminal pro‐B‐type natriuretic peptide (per 100 pg/mL; HR 1.02, 95% CI 1.01–1.03; P < 0.001), and rest scar burden (HR 1.03, 95% CI 1.01–1.06; P = 0.008) were independent risk factors for patients with AHF. The cross‐validated AUCs (95% CI) of nomogram constructed by diastolic blood pressure, valvular heart disease, cardiac resynchronization therapy, N‐terminal pro‐B‐type natriuretic peptide, and rest scar burden were 0.88 (0.73–1.00), 0.83 (0.70–0.97), and 0.79 (0.62–0.95) at 1, 3, and 5 years, respectively. Continuous net reclassification improvement and integrated discrimination improvement were also observed, and the decision curve analysis identified the greater net benefit of the nomogram across a wide range of threshold probabilities (0–100% at 1 and 3 years; 0–61% and 62–100% at 5 years) compared with dismissing the included factors or using either factor alone.ConclusionsA predictive nomogram for the risk of all‐cause mortality in patients with AHF was developed and validated in this study. The nomogram incorporated the rest scar burden by MPI is highly predictive, and may help to better stratify clinical risk and guide treatment decisions in patients with AHF.