Quantitative assessment of pleural effusion in critically ill patients by means of ultrasonography

Objective. To assess the ability of ultrasonography to identify the presence and assess the volume of pleural effusion in the intensive care unit setting. Design: Prospective descriptive clinical study. Setting. Medical-surgical intensive care unit of a teaching hospital. Patients. Initial study group (group 1) consisted of 97 patients (mean [+/- SD] Simplified Acute Physiology Score II, 40 +/- 14) with clinically suspected pleural effusion. Fifty-one patients were mechanically ventilated and 55 patients underwent a unilateral or bilateral thoracentesis (58 procedures). All patients underwent supine chest radiography and pleural ultrasonography at bedside. The testing group (group II) consisted of 19 additional patients (17 under ventilation) who underwent thoracentesis. Interventions: None. Measurements and Main Results: Maximal interpleural distance was measured at the base and apex of the pleural space, at both end-expiration and end-inspiration. In group 1, interpleural distances were compared to actual volume of fluid in the subset of patients who underwent a complete thoracentesis (n = 49). Prediction of the volume of pleural effusion was subsequently tested prospectively in group 11 (25 complete thoracenteses). Portable chest radiography and pleural ultrasonography yielded discordant results for 47 patients (48%) in the diagnosis of pleural effusion. The expiratory interpleural distance measured at the thoracic base with ultrasonography was significantly correlated with the volume of fluid (p < .0001; coefficient of determination: right, 0.78; left, 0.51). A pleural effusion >= 800 mL was predicted when this distance was > 45 mm (right) or > 50 mm (left), with a sensitivity of 94% and 100% and a specificity of 76% and 67%, respectively. In group II, the mean bias between the predicted and observed volumes of pleural effusion determined by thoracentesis was 24 +/- 355 mL, and this decreased to 28 +/- 146 mL for the prediction of pleural effusion < 1400 mL. Conclusions. Bedside ultrasonography is well suited for the quantitative assessment of unloculated pleural effusions in intensive care unit patients.