Abdominal sonography in examination of children with blunt abdominal trauma

OBJECTIVE: The objective of our study was to evaluate abdominal sonography for the detection of fluid and organ injury in children with blunt abdominal trauma. SUBJECTS AND METHODS: Fifty-one consecutive children with blunt abdominal trauma requiring abdominal CT were prospectively examined with sonography. Sonograms and CTs were independently evaluated by two radiologists for fluid and organ injury; CT examinations were considered abnormal if either was identified. Differences in CT interpretation were settled by a third observer. Using CT as the truth standard, we calculated the sensitivity, specificity, and negative predictive value of sonography for both observers. Agreement of the sonographic interpretations was evaluated using kappa statistic. RESULTS: In 33.3% of patients, CT revealed fluid, organ injury, or both. The sensitivity and specificity of sonography when detection of fluid was the sole parameter evaluated was 58.8% and 79.4%, respectively, for observer 1 and 47.1% and 79.4%, respectively, for observer 2. In contrast, the sensitivity and specificity of sonography when detection of both fluid and organ injury was evaluated was 64.7% and 79.4%, respectively, for observer 1 and 70.6% and 70.6%, respectively, for observer 2. The negative predictive value of sonography was 79.4% and 75.0% with evaluation limited to detection of fluid and 81.8% and 82.8% with evaluation of fluid and organ abnormality for observers 1 and 2 , respectively. Agreement was excellent for sonographic identification of fluid (kappa = 0.82) but poor for detection of organ injury (kappa = 0.34). CONCLUSION: The low sensitivity and negative predictive value of sonography when assessing for either fluid alone or fluid and organ injury suggest that a normal screening sonography alone in the setting of blunt abdominal trauma fails to confidently exclude the presence of an intraabdominal injury.