Increased utilization of computed tomography (CT) has led to a rise in the diagnosis of pulmonary contusion. Its clinical significance, in the absence of findings on chest radiograph (CXR), has not been defined. This study examines the clinical course of patients with CT-only diagnosis of pulmonary contusion and compares it with that of patients with CXR-proven pulmonary contusion. The trauma database identified all children undergoing chest CT for blunt thoracic trauma during a 3-year period. Records were reviewed for age, mechanism of injury, Injury Severity Score (ISS), length of hospital stay (LOS), need for intensive care unit admission, and need for endotracheal intubation. A pediatric radiologist reviewed all films in a blinded fashion. Statistical analysis was performed using analysis of variance and Fisher's Exact test for 2 x 3 tables. Eighty-two patients were identified. There were no CXR-positive, CT-negative cases. A CT diagnosis of pulmonary contusion was made in 46 patients. Of these, 31 had a contusion on CXR as well (CXR+ group) and 15 had a normal CXR (CT+ only group). Mean ISS score did not differ significantly between the two groups (27 +/- 12.3 and 22 +/- 10.3, respectively). Thirty-six patients had a normal CT (control). Mean LOS was significantly longer in the CXR+ group (13 +/- 12.0 days) than in the CT+ only and control groups (5 +/- 3.6 and 9 +/- 9.5 days, respectively; P < .01). The percentages of children requiring intensive care unit admission and intubation were also significantly higher in the CXR+ group. The finding of pulmonary contusion by CT alone does not increase patient morbidity and appears to be of limited clinical significance.