Objectives N-terminal pro-brain natriuretic peptide (NT-proBNP) is widely used in the diagnosis and prognostic assessment of heart failure. However, the possible influence of atrial fibrillation (AF) on NT-proBNP in the diagnosis of heart failure is still a matter of controversy. We analyse the influence of AF on NT-proBNP levels in dyspneic patients suspected with acute heart failure (AHF).
Methods 457 patients with dyspnea as the most prominent symptom and obtained a blood NT-proBNP measurement and electrocardiogram within 24 hours of hospitalisation were enrolled. The diagnosis of AF was identified by ECG.A final diagnosis of AHF was made by two advanced cardiologists according to Framingham heart failure criteria in combination with all hospital records. All involved patients were divided into AF or non-AF group according to ECG results, and also AHF or non-AHF group according to the diagnosis. NT-proBNP levels were compared between AF group and non-AF group in all dyspneic subjects, also in AHF patients and in non-AHF patients respectively. ROC analyses were performed to evaluate the diagnostic value of plasma NT-proBNP for AHF in both AF group and non-AF group, and to identify the optimal cut-off point for diagnosing AHF in these two groups.
Results Of 457 dyspneic patients (male 49.3%; mean age: 66.7 ± 14.9 years), 190 (42.5%) were diagnosed with AHF, 140 (13%) with AF. The median NT-proBNP level in AF group was significantly higher than in non-AF group (4482.0 pg/ml vs 1302.0 pg/ml, P = 0.000). 46.9% patients showed AF in AHF group. However, there was no difference of NT-proBNP levels between AHF patients with and without AF (6580.0 pg/ml vs 6769.0 pg/ml, P = 0.886). 18.6% patients showed AF in non-AHF group. The median NT-proBNP concentrations were significantly higher in non-AHF with AF patients than those in non-AHF without AF patients (2892.0 pg/ml vs 403.3 pg/ml, P = 0.000). Diagnostic accuracy of NT-proBNP for AHF was lower in AF patients than that in non-AF patients [AUC: 0.759, 95%CI (0.677-0.841) vs 0.931, 95%CI (0.903-0.985), P < 0.050]. The optimal cut-point of NT-proBNP for diagnosing the AHF was 3700 pg/ml in AF patients, with a sensitivity of 74%, specificity of 65%; and 1900 pg/ml in non-AF ptients, with a sensitivity of 91%, specificity of 79%.
Conclusions As atrial fibrillation is associated with higher NT-proBNP concentrations in dyspneic patients without AHF, it may influence the utility of NT-proBNP in the diagnosis of acute heart failure.