Impact of B-type natriuretic peptide level on the risk of left atrial appendage thrombus in patients with nonvalvular atrial fibrillation: a prospective study

BACKGROUND The impact of B-type natriuretic peptide (BNP) level on the risk of left atrial appendage (LAA) thrombus in patients with nonvalvular atrial fibrillation (NVAF) has not been prospectively studied. METHODS In two academic medical centers, we obtained BNP levels immediately prior to transesophageal echocardiogram performed to exclude LAA thrombus in patients with NVAF. RESULTS Among 261 subjects (mean age 65 ± 12 years; 30 % women) with NVAF, 17 (6.5 %) had LAA thrombus and 85 (32.6 %) had at least mild spontaneous echo contrast (SEC). Mean BNP level was significantly higher in patients with LAA thrombus [775 ± 678 vs. 384 ± 537, P = 0.001]. Receiver operator characteristics analysis demonstrated that BNP has a good discriminatory capacity for LAA thrombus (area under the curve, 0.74; 95 % confidence interval [CI], 0.63-0.85; P = 0.001); BNP ≥ 67 pg/mL was 100 % sensitive and 20 % specific for LAA thrombus. Multivariate logistic regression analysis demonstrated that BNP was not independently associated with LAA thrombus (odds-ratio, 1.05 per 100 pg/mL increment; CI, 0.99-1.12; P = 0.127) after adjusting for CHA2DS2-VASc score; while the latter was independently associated with LAA thrombus after adjusting for BNP level (odds-ratio, 1.46 per CHA2DS2-VASc point; CI, 1.09-1.96; P = 0.011). Nonetheless, BNP was associated with SEC in univariate and multivariate analysis, after adjusting for the CHA2DS2-VASc score, (odds-ratio, 1.08; CI, 1.02-1.14; P = 0.005). CONCLUSIONS BNP is predictive of SEC. However, it does not provide significant incremental value in the prediction of LAA thrombus.