Abstract
This study aims to explore the effect of N-terminal pro-brain natriuretic peptide (NT-proBNP) variability (mean absolute difference of the log2 NT-proBNP level measured in hospital) on the prognosis of patients with cardiorenal syndrome (CRS) type 2. Patients with CRS type 2 were retrospectively included. The varied NT-proBNP indications were analyzed. They were NT-proBNP I(pre-treatment), NT-proBNP II(post-treatment), NT-proBNP II/I, ΔNT-proBNP, log2 (NT-proBNP) variability and mean log2 (NT-proBNP). A logistic regression model and survival curves (Kaplan-Meier analysis) were built to identify independent predictors associated with poor prognosis. The primary outcomes were major adverse renal and cardiac events. The secondary outcome was all-cause mortality. From 2012 to 2016, 136 patients were included in this study with 69 (50.7%) had high log2 (NT-proBNP) variability level. The optimal cutoff level for each NT-proBNP indication that predicts poor prognosis was calculated, and the area under curves ranged from 0.668 to 0.891 with different indications. Kaplan-Meier analysis revealed that there was significantly correlated with prevalence of primary outcomes and NT-proBNP variability. The hazard ratios (HRs) ranged from 1.67 to 6.61 with different indications. The multivariate regression analyses also identified the risk of the primary outcomes were associated with elevated NT-proBNP values, except NT-proBNP I. The odds ratio (ORs) ranged from 1.83 to 6.61 with different indications. When analyzing the relationship between NT-proBNP variability and all-cause mortality, the results were the same. NT-proBNP variability might serve as an independent predictor for poor prognosis and all-cause mortality in patients with CRS type 2.