Assessment of a bedside test for N-terminal pro B-type natriuretic peptide (NT-proBNP) to differentiate cardiac from non-cardiac causes of pleural effusion in cats

BACKGROUND Cats with pleural effusion represent common emergencies in small animal practice. The aim of this prospective study was to investigate the diagnostic ability of a point-of-care ELISA (POC-ELISA) for the measurement of N-terminal pro B-type natriuretic peptide (NT-proBNP) to differentiate cardiac from non-cardiac disease in cats with pleural effusion. The sample material for use of this rapid test was either plasma or diluted pleural effusion. Twenty cats with moderate to severe pleural effusion were prospectively recruited. The cats were grouped into two groups, with or without congestive heart failure (CHF; N-CHF), after complete work-up. Blood and effusion were collected in EDTA tubes. Plasma and pleural effusion supernatants were transferred into stabilizer tubes and frozen. POC-ELISA for NT-proBNP was performed with plasma and diluted effusion (1:1). Quantitative NT-proBNP measurement was performed in plasma and diluted and undiluted effusions. RESULTS Six cats were assigned to the CHF group. Of the 14 cats in the N-CHF group, 6 had concurrent cardiac abnormalities that were not responsible for the effusion. For the detection of CHF, the test displayed respective sensitivities and specificities of 100% and 79% in plasma and 100% and 86% in diluted pleural fluid. Receiver operating characteristic (ROC) analysis for quantitative NT-proBNP measurement of plasma and diluted and undiluted pleural effusions displayed areas under the curve of 0.98, sensitivities of 100% and specificities of 86%. The optimum cut-off was calculated at 399 pmol/l in plasma and 229 pmol/l in the diluted effusion and 467 pmol/l in the undiluted effusion. CONCLUSIONS POC-ELISA for NT-proBNP in both plasma and diluted pleural effusion was suitable to differentiate cardiac from non-cardiac causes of feline pleural effusion. According to our results, use of pleural effusion is feasible, but dilution of the effusion before measurement seems to improve specificity.