A pneumothorax is a potentially life-threatening condition. Although CT scan is the reference standard for diagnosis, chest radiographs are commonly used to rule out the diagnosis. We compared the test characteristics of ultrasonography and supine chest radiography in adult patients clinically suspected of having a pneumothorax, using CT scan or release of air on chest tube placement as reference standard.
We searched for English literature in MEDLINE and EMBASE and performed hand searches. Two independent investigators used standardized forms to review articles for inclusion, quality (QUADAS tool), and data extraction. We calculated κ agreement for study selection and evaluated clinical and quality homogeneity before meta-analysis.
We reviewed 570 articles and selected 21 for full review (κ, 0.89); eight articles (total of 1,048 patients) met all inclusion criteria (κ, 0.81). All studies but one used the ultrasonographic signs of lung sliding and comet tail to rule out pneumothorax. Chest radiography data were available for 864 of 1,048 patients evaluated with ultrasonography. Ultrasonography was 90.9% sensitive (95% CI, 86.5-93.9) and 98.2% specific (95% CI, 97.0-99.0) for the detection of pneumothorax. Chest radiography was 50.2% sensitive (95% CI, 43.5-57.0) and 99.4% specific (95% CI, 98.3-99.8).
Performance of ultrasonography for the detection of pneumothorax is excellent and is superior to supine chest radiography. Considering the rapid access to bedside ultrasonography and the excellent performance of this simple test, this study supports the routine use of ultrasonography for the detection of pneumothorax.