BACKGROUND: Definite diagnosis of transudative or exudative pleural fluids often presents a diagnostic dilemma. The aim of this study was to evaluate whether amino-terminal brain natriuretic peptide (NT-proBNP) levels in pleural fluid has a diagnostic value for discriminating heart-failure-related pleural effusions from non-heart-failure effusions. METHODS: Sixty-six subjects (40 male, mean age 61 ± 18 y) with pleural effusions were included. Samples of pleural fluid and serum were obtained simultaneously from each subject. Biochemical analysis, bacterial and fungal culture, acid-fast bacilli smear and culture, and cytology were performed on the pleural fluid. RESULTS: Subjects with heart-failure-related pleural effusion had significantly higher pleural NT-proBNP levels than other subjects (P < .001). Pleural and serum NT-proBNP measures were closely correlated (r = 0.90, P < .001). An NT-proBNP cutoff value of ≥ 2,300 pg/mL in pleural fluid had a sensitivity of 70.8%, a specificity of 97.6%, and positive and negative predictive values of 94.4% and 85.4%, respectively, for discriminating transudates caused by heart failure from exudates. Eight heart-failure subjects were misclassified as exudates by Light's criteria, 5 of whom received diuretics before thoracentesis. All misclassified subjects had pleural NT-proBNP levels higher than 1,165 pg/mL, which predicted heart-failure-associated transudates with 95.8% sensitivity and 85.7% specificity. CONCLUSIONS: Pleural fluid NT-proBNP measurement in the routine diagnostic panel may be useful in differentiation of heart-failure-related pleural effusions and exudative pleural fluids with reasonable accuracy, especially in heart-failure patients treated with diuretics.
BACKGROUND: Definite diagnosis of transudative or exudative pleural fluids often presents a diagnostic dilemma. The aim of this study was to evaluate whether amino-terminal brain natriuretic peptide (NT-proBNP) levels in pleural fluid has a diagnostic value for discriminating heart-failure-related pleural effusions from non-heart-failure effusions. METHODS: Sixty-six subjects (40 male, mean age 61 ± 18 y) with pleural effusions were included. Samples of pleural fluid and serum were obtained simultaneously from each subject. Biochemical analysis, bacterial and fungal culture, acid-fast bacilli smear and culture, and cytology were performed on the pleural fluid. RESULTS: Subjects with heart-failure-related pleural effusion had significantly higher pleural NT-proBNP levels than other subjects (P < .001). Pleural and serum NT-proBNP measures were closely correlated (r = 0.90, P < .001). An NT-proBNP cutoff value of ≥ 2,300 pg/mL in pleural fluid had a sensitivity of 70.8%, a specificity of 97.6%, and positive and negative predictive values of 94.4% and 85.4%, respectively, for discriminating transudates caused by heart failure from exudates. Eight heart-failure subjects were misclassified as exudates by Light's criteria, 5 of whom received diuretics before thoracentesis. All misclassified subjects had pleural NT-proBNP levels higher than 1,165 pg/mL, which predicted heart-failure-associated transudates with 95.8% sensitivity and 85.7% specificity. CONCLUSIONS:Pleural fluid NT-proBNP measurement in the routine diagnostic panel may be useful in differentiation of heart-failure-related pleural effusions and exudative pleural fluids with reasonable accuracy, especially in heart-failurepatients treated with diuretics.