The aim of therapeutic thoracentesis (TT) is to aspirate as much pleural fluid as possible. Monitoring pleural pressure (PlP) during TT has been proposed to avoid the adverse effects due to an unintended sharp drop in PlP. The objectives of this study are to ascertain the diagnostic value of the PlP measurement, to find a predictive variable of the amount of fluid that can be removed, to obtain insight into the characteristics of the PlP curve and pleural elastance (PE) during TT, and to describe the complications of TT. Sixty-one unselected patients were studied. Only the four patients with suspected trapped lung had an initial PlP lower than -4 cm H(2)O and a PE higher than 33 cm H(2)O/L. There was a weak correlation (r = 0.52) between PE during the first 0.5 L aspirated and the total amount of fluid aspirated. Partial PE values were 10, 7.5, and 14 cm H(2)O/L at the early, intermediate, and late phases of TT. No complications were found except for nine pneumothoraces. In conclusion, the technique was clinically helpful because large amounts of pleural fluid could be aspirated with few and mild complications, and because it allows clinicians to support the preliminary diagnosis of trapped lung. None of the studied variables was found to predict the suitability of aspirating more than 1.5 L. Rather than being monotonically descendent, the PlP curve shows a three-part line with the deepest slopes at the first and last phases of the thoracentesis.