Predictors of "occult" intra-abdominal injuries in blunt trauma patients

to assess predictors of intra-abdominal injuries in blunt trauma patients admitted without abdominal pain or abnormalities on the abdomen physical examination.

We conducted a retrospective analysis of trauma registry data, including adult blunt trauma patients admitted from 2008 to 2010 who sustained no abdominal pain or abnormalities on physical examination of the abdomen at admission and were submitted to computed tomography of the abdomen and/or exploratory laparotomy. Patients were assigned into: Group 1 (with intra-abdominal injuries) or Group 2 (without intra-abdominal injuries). Variables were compared between groups to identify those significantly associated with the presence of intra-abdominal injuries, adopting p<0.05 as significant. Subsequently, the variables with p<0.20 on bivariate analysis were selected to create a logistic regression model using the forward stepwise method.

A total of 268 cases met the inclusion criteria. Patients in Group I were characterized as having significantly (p<0.05) lower mean AIS score for the head segment (1.0 ± 1.4 vs. 1.8 ± 1.9), as well as higher mean AIS thorax score (1.6 ± 1.7 vs. 0.9 ± 1.5) and ISS (25.7 ± 14.5 vs. 17,1 ± 13,1). The rate of abdominal injuries was significantly higher in run-over pedestrians (37.3%) and in motorcyclists (36.0%) (p<0.001). The resultant logistic regression model provided 73.5% accuracy for identifying abdominal injuries. The variables included were: motorcyclist accident as trauma mechanism (p<0.001 - OR 5.51; 95%CI 2.40-12.64), presence of rib fractures (p<0.003 - OR 3.00; 95%CI 1.47-6.14), run-over pedestrian as trauma mechanism (p=0.008 - OR 2.85; 95%CI 1.13-6.22) and abnormal neurological physical exam at admission (p=0.015 - OR 0.44; 95%CI 0.22-0.85).

Intra-abdominal injuries were predominantly associated with trauma mechanism and presence of chest injuries.