Lung contusion in children--early computed tomography versus radiography.

ABSTRACT

OBJECTIVE: To investigate, in children, the correlation between the extent of lung contusion as detected on early radiologic examination (chest radiograph [CXR] and/or thoracic computed tomography [TCT]) and subsequent clinical outcome measures. DESIGN: Retrospective chart review study with blinded assessment of thoracic imaging. SETTING: A university-affiliated, level 1 designated pediatric trauma center. INTERVENTIONS: None. PATIENTS Patients (1-18 yrs) who, between April 2000 and October 2005, were diagnosed with lung contusion were eligible for study entry. The medical records of those patients who underwent early (within the first 24 hrs of admission) thoracic imaging (CXR and/or TCT) were reviewed. A pulmonary contusion score (PCS) was assigned to each thoracic image according to the extent of contusion injury by two investigators blinded to each others score and the clinical details of the patient. RESULTS: Seventy-four patients were included in the study. Twenty patients had undergone CXR only, whereas 54 had undergone both CXR and TCT. The mean PCS on CXR was 3.9 +/- 3.6 compared with 6.5 +/- 3.49 on TCT (p < .001). In eight patients (15%) who underwent TCT and CXR, the CXR failed to demonstrate a lung contusion. The PCS derived from CXR examination correlated positively with lower Pao2/Fio2 (r = -.36, p = .019), higher ventilation index (r = .35, p = .014), and longer length of ventilation (r = .28, p = .019). No such correlation was seen with TCT-derived PCS. CONCLUSIONS: The severity of lung contusion determined by CXR, but not TCT, correlates with impairment of oxygenation, CO2 exchange, and duration of ventilatory support.