Background: Pulmonary contusion (PC) is a common injury associated with blunt chest trauma. Complications such as pneumonia and adult respiratory distress syndrome (ARDS) occur in up to 50% of patients with PC. The ability to predict which PC patients are at increased risk of developing complications would be of tremendous clinical utility. In this study, we test the hypothesis that a novel method that objectively measures percent PC can be used to identify patients at risk to develop ARDS after injury.
Methods: Patients with unilateral or bilateral PC with an admission chest computed tomographic angiogram were identified from the trauma registry. Demographic, infectious, and outcome data were collected. Percent PC was determined on admission chest computed tomography using our novel semiautomated, attenuation-defined computer-based algorithm, in which the lung was segmented with minimal manual editing. Factors contributing to the development of ARDS were identified by both univariate and multivariable logistic regression analyses. ARDS was defined as PaO2/FiO2 ratio of less than 200 with diffuse bilateral infiltrates on chest radiograph with no evidence of congestive heart failure.
Results: Quantifying percent PC from our objective computer-based approach proved successful. We found that a contusion size of 24% of total lung volume or greater was most significant at predicting ARDS, which occurred in 78% of these patients. Such patients also had a significantly higher incidence of pneumonia when compared with those with contusions less than 24%. The specificity of contusion size of 24% or greater was 94%, although sensitivity was 37%; positive predictive value was 78%, and negative predictive value was 72%.
Conclusion: We developed and describe a software-based methodology to accurately measure the size of lung contusion in patients of blunt trauma. In our analyses, contusions of 24% or greater most significantly predict the development of ARDS. Such an objective approach can identify patients with PC who are at increased risk for developing respiratory complications before they happen. Further research is needed to use this novel methodology as a means to prevent posttraumatic lung injury in patients with blunt trauma.
Level of evidence: Prognostic/epidemiologic study, level III; diagnostic study, level IV.