P63 A prospective observational trial examining the diagnostic utility of serum and pleural fluid procalcitonin in the initial investigation of unilateral pleural effusions

Differentiation of pleural infection from other causes of pleural effusion, particularly pleural malignancy can be difficult. Clinical features and readily available tests are non-specific resulting in early over-diagnosis of pleural infection. Procalcitonin (PCT) is produced in response to acute bacterial infection and its measurement in serum has shown promise in directing and shortening antibiotic courses in lower respiratory tract infections. We prospectively examined the diagnostic accuracy of PCT in the investigation of unilateral pleural effusions. Methods Consecutive patients with a unilateral pleural effusion, referred to a UK teaching hospital were included. Baseline serum and pleural fluid PCT was measured by enzyme-linked-fluorescent-assay (Vidas BRAHMS PCT—BioMerieux) and results compared to final diagnosis. Clinical data were collected prospectively and the ultimate diagnosis agreed against established criteria by two respiratory consultants, blind to the PCT result. Patients were followed up to histological or microbiological diagnosis, radiographic resolution or for 12 months. Results 145 patients, median age 72 (31–96). 71 inpatients, 75 outpatients. Effusion diagnoses: Complicated parapneumonic (CPE) 26/145, Simple parapneumonic (SPE) 7/145, Malignant 73/145, Idiopathic pleuritis 4/145, TB 1/145, other benign cause 34/145. Four patients with symptomatic non-thoracic bacterial infection and four with frank empyema were excluded from analysis. The Receiver operating characteristic (ROC) curve for serum PCT distinguishing CPE from non-infective diagnoses gave an AUC of 0.779 (95% CI 0.658 to 0.899) and at an optimum cut-off of 0.09 ng/ml had sensitivity 73%, specificity 81%, PPV 44% and NPV 94%. The ROC curve for pleural PCT gave an AUC of 0.809 (95% CI 0.709 to 0.908) and at an optimum cut-off of 0.1 ng/ml had sensitivity 81%, specificity 78%, PPV 45% and NPV 95%. Sub-group analysis excluding patients with >48 h in patient stay or >48 h antibiotics further improved diagnostic accuracy. In patients with pleural fluid pH ≤7.2 serum PCT distinguished CPE from other effusions with an AUC of 0.808 (0.613–1.000). Conclusion Serum and pleural fluid procalcitonin have promising diagnostic characteristics in distinguishing complicated parapneumonic effusions (Abstract P63 Table 1) from unilateral effusions of non-infective origin, having particularly high negative predictive values. Procalcitonin could help the clinician to decide optimal management pathways for individual patients.