Background: Ultrasound has proven to be very accurate in the diagnosis of pneumothorax in the trauma suite. It is unknown whether this accuracy is maintained over time in patients with a thoracostomy (TT) in place.
Methods: Hospitalized patients with a TT placed to treat a traumatic pneumothorax underwent serial daily bedside surgeon-performed ultrasound by 1 of 2 experienced surgeon sonographers who were unaware of concomitant X-ray findings. Results were compared with daily chest X-ray films. Data collected included size and day of placement of the chest tube, as well as the results of the serial ultrasounds and the comparative X-ray films.
Results: Fourteen patients (78% men, mean age 33 years) sustained traumatic pneumothorax. The causes included stab wound (9), gunshot wound (3), and rib fracture (2). They underwent 126 (median 7) ultrasound evaluations and were followed between 4 and 26 (median 7) days after injury. Of these exams, 95 had a concomitant chest X-ray film within 1 hour of the ultrasound, thus 190 hemithoraces could be analyzed. Eighty-two ultrasounds were performed for hemithoraces that had no injury or TT in place and all 82 revealed normal pleural sliding. No pneumothoraces were noted on concomitant chest X-ray films (100% accuracy). One hundred eight ultrasounds were performed for hemithoraces that had a TT in place. For the first 24 hours, accuracy remained 100%. After 24 hours, however, sensitivity of ultrasound diagnosis of pneumothorax fell to 55%, specificity fell to 70%, positive predictive value to 43%, and negative predictive value to 79%. This led to an overall accuracy rate for ultrasound examination after 24 hours of 65%.
Conclusions: Ultrasound evaluation for pneumothorax is very accurate for the first 24 hours after insertion of a TT, but the accuracy, especially the positive predictive value, is not sustained over time, possibly as a result of the formation of intrapleural adhesions.