Background and objective: Parapneumonic effusions with pH < 7.20 or glucose < 0.40 g/l or lactate dehydrogenase (LDH) > 1000 U/l have indication of treatment with endothoracic drainage tube (EDT). The aim of the present study was to determine the accuracy of partial pressure of carbon dioxide (pCO2) measurement in pleural fluid for the subsequent treatment indication with EDT in parapneumonic effusions, by analyzing the area under curve ROC (AUC) and determining the optimal cut off value.
Patients and method: 207 pleural fluids were studied. Glucose, LDH, pCO2, and pH were measured, and data concerning the etiology of pleural effusion and whether EDT treatment was needed were collected after patients were discharged from hospital.
Results: Forty-six out of 207 pleural fluids studied were parapneumonic effusions. Thirty-two were treated with EDT. AUC values were 0.888 (p < 0.0001), 0.890 (p < 0.0001), 0.816 (p < 0.0001), and 0.801 (p < 0.0001) for pCO2, pH, glucose, and LDH, respectively. No significant differences were found among them. Optimal cut off value for pCO2 was 48.6 mmHg, exhibiting 90.6% sensitivity and 78.6% specificity. All parapneumonic effusions showing pCO2 > 60.9 mmHg were treated with EDT. Remarkably, 3 out of 46 parapneumonic effusions (6.5%) that had been improperly treated following pH, glucose or LDH values, were correctly treated following pCO2.
Conclusions: pCO2 determination in pleural fluid appears to be the best way to decide the indication of EDT in parapneumonic effusions.