Comparing serum and pleural fluid pro-brain natriuretic peptide (NT-proBNP) levels with pleural-to-serum albumin gradient for the identification of cardiac effusions misclassified by Light's criteria.

BACKGROUND AND OBJECTIVES: To assess the diagnostic performance of the amino-terminal fragment of pro-brain natriuretic peptide (NT-proBNP) in pleural fluid and serum for the identification of pleural effusions owing to heart failure, and to determine if these measurements allow better categorization of cardiac effusions that have been misclassified by Light's criteria, than do serum-pleural fluid albumin and protein gradients.

METHODS: The study prospectively evaluated NT-proBNP in serum and pleural fluid from patients with effusions owing to heart failure (n = 53) and other causes (n = 40). Measurements were made of levels of NT-proBNP by an electrochemiluminiscence immunoassay, and serum-pleural fluid protein and albumin gradients.

RESULTS: Using a cut-off value of 1500 pg/mL for serum and pleural samples, the accuracy of NT-proBNP for identifying pleural effusions from cardiac causes was 89% and 90%, respectively. The area under the receiver operating characteristic curve for the diagnosis of pleural effusions from heart failure was similar for pleural fluid (0.931, 95% CI: 0.871-0.991) and serum (0.919, 95% CI: 0.855-0.984) NT-proBNP. Six (75%) of eight patients with cardiac effusions that were misclassified as exudates by Light's criteria would have been correctly categorized by either NT-proBNP or the albumin gradient, whereas only four (50%) would have been correctly classified by the protein gradient.

CONCLUSIONS: NT-proBNP is a useful marker for the diagnosis of pleural effusions from heart failure when measured in either serum or pleural fluid. At a cut-off of 1500 pg/mL, NT-proBNP is at least as accurate as the albumin gradient to correctly identify cardiac effusions misclassified as exudates by standard criteria, but at much higher cost.