Background: The optimal volume of pleural fluid to diagnose a malignant effusion is unknown. Our study was designed to demonstrate if a minimum pleural fluid volume (10 mL) is equivalent to a large volume thoracentesis to make a cytopathologic diagnosis of malignancy.
Methods: A total of 121 thoracentesis samples were obtained from 102 patients with suspected or known malignant effusions. Pleural fluid was collected in three aliquots for cytologic examination (10 mL, 60 mL, > or = 150 mL). The pathologist was blinded to patient identifiers and aliquot volume. Sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) were calculated for each volume for the diagnosis of malignancy.
Results: Pleural malignancy was diagnosed in 90 patient encounters (74.4%). For direct smear/cytospin, there was increased sensitivity and NPV for 60 mL (P = .0058 and P = .045, respectively) and for > or = 150 mL (P < .001 and P = .009, respectively) compared with 10 mL. For combined direct smear/cytospin and cell block preparations, statistical significance for sensitivity and NPV existed only between the 10 mL and > or = 150 mL specimens (P = .0099 and P = .033, respectively). No statistical difference existed for specificity or PPV for any aliquot volume.
Conclusions: The sensitivity for diagnosis of pleural malignancy is dependent on the pleural fluid volume extracted during thoracentesis. Volumes of 10 mL do not perform as well as larger volumes. When both direct smear/cytospin and cell block preparations are used, we recommend > or = 150 mL, whereas when only direct smear/cytospin is used, 60 mL is adequate for the diagnosis a malignant pleural effusion.