Criteria for the selective use of chest computed tomography in blunt trauma patients.

Abstract

The purpose of this study was to derive parameters that predict which high-energy blunt trauma patients should undergo computed tomography (CT) for detection of chest injury. This observational study prospectively included consecutive patients (>or=16 years old) who underwent multidetector CT of the chest after a high-energy mechanism of blunt trauma in one trauma centre. We included 1,047 patients (median age, 37; 70% male), of whom 508 had chest injuries identified by CT. Using logistic regression, we identified nine predictors of chest injury presence on CT (age >or=55 years, abnormal chest physical examination, altered sensorium, abnormal thoracic spine physical examination, abnormal chest conventional radiography (CR), abnormal thoracic spine CR, abnormal pelvic CR or abdominal ultrasound, base excess <-3 mmol/l and haemoglobin <6 mmol/l). Of 855 patients with >or=1 positive predictors, 484 had injury on CT (95% of all 508 patients with injury). Of all 192 patients with no positive predictor, 24 (13%) had chest injury, of whom 4 (2%) had injuries that were considered clinically relevant. Omission of CT in patients without any positive predictor could reduce imaging frequency by 18%, while most clinically relevant chest injuries remain adequately detected.