Computed tomography-based risk estimation on acute lung injury/acute respiratory distress syndrome after blunt thoracic trauma

Objective

Computed tomography (CT) is used in the diagnostic management of polytraumatized patients. Multiple trauma and pulmonary contusion are typical triggers of post-traumatic acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) [1]. Therefore, an early predictor of post-traumatic ALI/ARDS would be valuable. In this study we tested whether the mass of nonaerated lung tissue (Mnon) in admission CT could predict post-traumatic ALI/ARDS.

Methods

The Mnon of 54 polytraumatized patients with pulmonary contusion was analyzed as previously described [2]. We studied the association of Mnon with physiologic variables and injury descriptors such as PaO₂/FiO₂ ratio, injury severity score (ISS) and thoracic trauma severity score (TTSS) [3] recorded on admission. To evaluate Mnon as a predictor of ALI/ARDS we used a receiver operator characteristic (ROC) curve.

Results

Patients developing post-traumatic ALI/ARDS had significantly larger Mnon and significantly higher ISS and TTSS values. Literature data suggest a higher incidence of ALI/ARDS with a contusioned lung volume >20% [4]. Our results, however, indicate a higher risk of ALI/ARDS already with a Mnon of around 10%. The Mnon predicting ALI/ARDS with the highest sensitivity (81%) and specificity (87%) was 9.8%. The area under the ROC curve was 0.89 (confidence interval 0.79–0.99).

Conclusion

The mass of nonaerated lung tissue on admission CT can help to predict the development of ALI/ARDS. It may thereby help to implement appropriate therapeutic options such as lung protective ventilation. The clinical use of our technique, however, is limited by the time-consuming CT analysis.