Occult traumatic pneumothorax: diagnostic accuracy of lung ultrasonography in the emergency department.

BACKGROUND The role of chest ultrasonography (US) in the diagnosis of pneumothorax (PTX) has been established, but how it compares with lung CT scanning in the diagnosis of radiooccult PTX and in the determination of its topographic extension has not yet been completely evaluated. OBJECTIVE To determine the diagnostic accuracy of chest US in the emergency department (ED) in the diagnosis of occult PTX in trauma patients and to define its ability to determine PTX extension. DESIGN An 18-month prospective study. PATIENTS A total of 109 conscious, spontaneously breathing patients who had been admitted to the ED for chest trauma or polytrauma. METHODS All eligible patients underwent a standard anteroposterior supine chest radiograph (Rx) and a spiral CT lung scan within 1 h of ED admission. Lung US was carried out by an operator who was unaware of the other examination results, both for diagnosis and for the quantitative delimitation of the PTX. RESULTS Twenty-five traumatic PTXs were detected in the 218 hemithoraxes (109 patients; 2 patients had a bilateral PTX) evaluated by spiral CT scan; of these, only 13 of 25 PTXs (52%) were revealed by chest Rx (sensitivity, 52%; specificity, 100%), while 23 of 25 PTXs (92%) were identified by lung US with one false-positive result (sensitivity, 92%; specificity, 99.4%). In 20 of 25 cases, there was agreement on the extension of the PTX between CT lung scan and lung US with a mean difference of 1.9 cm (range, 0 to 4.5 cm) in the localization of retroparietal air extension; chest Rx was not able to give quantitative results. CONCLUSIONS Lung US scans carried out in the ED detect occult PTX and its extension with an accuracy that is almost as high as the reference standard (CT scanning).