Brain natriuretic peptide is a biomarker of atrial fibrillation in hemodialysis patients

Abstract

Objectives: Atrial fibrillation (AF) is the most common persistent arrhythmia and has adverse outcomes in hemodialysis patients. It is obscure the relationship between BNP and AF in hemodialysis patients. This study investigated the interventionable predictors of AF in hemodialysis patients.

Methods: In this retrospectively observational cohort study, a total of 205 hemodialysis patients were screened and 140 were enrolled. The anthropometrics, laboratory parameters, electrocardiogram and echocardiogram were collected. The patients were divided into three groups based on the tertiles of BNP value. Kaplan-Meier curves were used to compare the incidence of AF among different BNP groups. Restricted cubic spline curves and receiver operator characteristic curves were drawn, and Cox proportional hazards models were applied to identify the predictive value of BNP and the other related factors for incident AF.

Results: During the 5-year follow-up period, 33 (23.6%) individuals developed incident AF. The incidence of AF increased significantly with an increase in the BNP. Cox proportional hazards models indicated that age, dialysis vintage, left atrial diameter, ultrafiltration rate, hs-CRP and BNP were independent risk factors for incident AF. The hazard ratios of BNP (per 100 pg/mL) were 1.038 (95% confidence interval, 1.012-1.064, p = 0.004). BNP had a predictive value for the occurrence of AF (area under the curve = 0.734).

Conclusions: BNP was a good predictor of incident AF in hemodialysis patients. Higher BNP had an increasing adverse event rate of AF. Further research should be needed to clarify the best reference range of BNP in hemodialysis patients.

Keywords: Atrial fibrillation; brain natriuretic peptide; hemodialysis; predictor; ultrafiltration rate.

Figure 1.

Flow diagram of recruited patients.

Figure 2.

The incidence rate of AF in the different BNP tertiles. The number of AF/total patients in each tertile was 5/46, 9/47 and 19/47, respectively (χ²= 12.041, p = 0.001).

Figure 3.

Kaplan–Meier Plots for the incidence rate of AF in the different BNP groups (log-rank test).

Figure 4.

Univariate associations between BNP and incident AF when BNP is modeled as a cubic spline. The solid blue line indicates HRs for incident AF as a function of BNP and the light gray shading represents the associated 95% CIs.

Figure 5.

The ROC curves of BNP, UR and their combining model in assessing incident AF. (BNP, AUC = 0.734, sensitivity = 0.758, specificity = 0.626; UR, AUC= 0.636, sensitivity = 0.848, specificity = 0.421; combining model, AUC= 0.740, sensitivity = 0.485, specificity = 0.879).