Anatomical Injury Clusters in Polytrauma Patients

Abstract

Polytrauma is a major cause of death in young adults. The trial was to identify clusters of interlinked anatomical regions to improve strategical operational planning in the acute situation. A total of 2219 polytrauma patients with an ISS (Injury Severity Score) ≥ 16 and an age ≥ 16 years was included into this retrospective cohort study. Pearson's correlation was performed amongst the AIS (Abbreviated Injury Scale) groups. The predictive quality was tested by ROC (Receiver Operating Curve) and their area under the curve. Independency was tested by the binary logistic regression, AIS ≥3 was taken as a significant injury. The analysis was conducted using IBM SPSS^® 24.0. The highest predictive value was reached in the combination of thorax, abdomen, pelvis and spine injuries (ROC: abdomen for thorax 0.647, thorax for abdomen 0.621, pelvis for thorax 0.608, pelvis for abdomen 0.651, spine for thorax 0.617). The binary logistic regression revealed the anatomical regions thorax, abdomen pelvis and spine as per-mutative independent predictors for each other when a particular injury exceeded the AIS ≥3. The documented clusters of injuries in truncal trauma are crucial to define priorities in the polytrauma management.

Keywords: AIS; ISS; Polytrauma; Retrospective Cohort Study; Truncal Trauma.

Figure 1:

Relative density estimation of the injury-clusters depicted by contour plots. The association between abdominal injuries and spinal injuries is evident (A). Only the highly scored abdominal and thoracic injuries led to a concomitant pelvic injury (B). Spinal injuries are broadly associated with thoracic injuries (C), and pelvic injuries with abdominal injuries. Data are given as artificial units representing relative density of the respective event.