Study objective: To compare the sensitivity, specificity, and accuracy of ultrasonography with those of the initial plain chest radiograph for detection of hemothorax in trauma patients.
Methods: Data from a prior prospective study of trauma ultrasonography at a Level I trauma center were retrospectively analyzed. The medical records of a convenience sample of adult patients who presented with major blunt or penetrating torso trauma during a 17-month period were reviewed. Emergency physicians performed a trauma ultrasound examination, which included evaluation for pleural fluid. Ultrasound interpretations were recorded before other diagnostic tests were obtained and were not used in patient management decisions. Records of the study patients were reviewed for confirmation of the presence or absence of hemothorax by other diagnostic and therapeutic interventions. The chest radiograph and computed tomography (CT) scan interpretations were performed by attending radiologists who were not blinded to patient outcome.
Results: Five of the 245 patients enrolled in the study were excluded because tube thoracostomy was performed before the ultrasound examination was done. Altogether, 26 of the 240 study patients had hemothorax, as confirmed by tube thoracostomy or CT. Both ultrasound examination and the initial chest radiograph resulted in 0 false-positive, 1 false-negative, 25 true-positive, and 214 true-negative findings. Overall, both modailties were 96.2% sensitive, 100% specific, and 99.6% accurate.
Conclusion: Ultrasonography is comparable to the initial chest radiograph for accuracy in detection of hemothorax and may expedite the diagnosis and treatment of this condition for patients with major trauma.