May the initial CT scan predict the occurrence of delayed hemothorax in blunt chest trauma patients?

PURPOSE: To assess the impact of delayed hemothorax on outcomes in blunt chest trauma patients without life-threatening condition at admission and characterize the predictive value of predefined anatomical factors for delayed hemothorax. METHODS: In a single-centre retrospective study, every spontaneous breathing patient admitted for a blunt chest trauma without significant pleural effusion at ICU admission was included. A multivariable regression model was used to determine the covariate-adjusted odd of secondary respiratory complications in patients with delayed hemothorax ≥ 500 ml. The characteristics of rib fractures (number, location and displacement) were integrated into a logistic regression model to determine variables associated with delayed hemothorax in multivariate analysis. RESULTS: Over the study period, 109 patients were included and the rate of delayed hemothorax ≥ 500 ml was 36%. Patients with delayed hemothorax had higher rates of pulmonary infections (OR 4.8 [1.6-16.4]) but no statistical association between delayed hemothorax and secondary respiratory failure (OR 2.0 [0.4-9.4]). A posterior location and a displaced rib fracture were independent predictors of delayed hemothorax (OR 3.4 [1.3-8.6] and OR 2.3 [1.1-5.1], respectively). At least one displaced rib fracture was more specific of delayed hemothorax than the commonly used threshold of three or more rib fractures (81.3 vs. 51.5%). CONCLUSION: Delayed hemothorax is a frequent complication associated with increased risk of pulmonary infection. The posterior location and the displacement of at least one rib fracture in the initial CT scan were independent risk factors for predicting the occurrence of delayed hemothorax.