P65 Is serum N-terminal pro B type natriuretic peptide (NT-proBNP) measurement useful in the investigation of unilateral pleural effusions?--a prospective observational study

Abstract

Measurement of serum NT-proBNP has been proposed in the investigation of pleural effusions, particularly in the diagnosis of cardiac failure in those misclassified as exudates by Light's criteria. Studies have reported excellent diagnostic accuracy for the test but have included both bilateral and unilateral effusions and applied short follow-up periods. We prospectively examined the diagnostic utility of serum NT-proBNP in a consecutive series of unilateral pleural effusions with robust follow-up and diagnostic criteria. Method(s): Consecutive patients presenting to a UK teaching hospital with an undiagnosed unilateral pleural effusion underwent clinical assessment including CXR, ECG, echocardiogram, thoracentesis (and CTwhen appropriate). Light's criteria were applied. Serum NTproBNP was measured using point of care ELISA. Patients were followed up to histological/microbiological diagnosis, radiographic resolution or 12 months. Echocardiograms were double reported and diagnosis determined independently by two respiratory consultantsdall blind to NT-proBNP results. Result(s): 118 patients. Median age 74 (42e95). 39 in patients, 79 outpatients. 18 transudates, 92 exudates. 30 large, 66 moderate, 22 small. Diagnoses: primary cardiac cause (PCC) 20/118, Malignant 57/118, PE 4/118, Non-cardiac transudate 3/118, other benign cause 30/118. The ROC curve for NT-proBNP discriminating effusions of PCC gave an AUC of 0.845 (0.774e0.934). At cut-off of age and sex adjusted 97.5th centile (healthy population) NT-proBNP had sensitivity 100%, Specificity 53%, PPV 30% and NPV 100% and all four cardiac exudates were correctly diagnosed. At an optimum cutoff of 1500 pg/mldsensitivity 75%, specificity 76%, PPV 38% and NPV 94%. Co-morbid cardiac disease was common in patients without a PCC for effusion with 70% having significant abnormalities on echocardiogram but cardiac disease was considered to be contributing to effusion in only 9/98 of this group. Conclusion(s): The excellent negative predictive value of NT-proBNP, particularly at an age and sex adjusted cut-off level gives the test utility to rule out a primarily cardiac cause in selected cases of unilateral pleural effusion. Co-morbid cardiac disease and associated NT-proBNP elevation is very common in patients with a noncardiac origin of pleural effusion such that a positive test at baseline should not alter the initial investigation pathway, particularly amongst pleural exudates.