Predictors of mortality in patients with isolated severe traumatic brain injury

Traumatic brain injury (TBI) is a leading cause of death and disability worldwide. Many prognostic models predicting mortality in patients with TBI were developed, which also include patients with mild or moderate TBI and patients who suffered major extracranial injuries. From a prospective database, we conducted a retrospective medical chart review covering the period between January 2000 and December 2012 of patients with isolated severe TBI (Abbreviated Injury Score for head, AIS(H) ae<yen> 3) without extracranial injuries, who were intubated in the field using the rapid sequence intubation method and were of age 16 or more. Prehospital vital signs, Injury Severity Score (ISS) and laboratory tests were compared in two study groups: survivors (n = 25) and non-survivors (n = 27). Selected variables identified during univariate analysis (p < 0.1) were then subjected to multivariate analysis logistic regression model. Univariate analysis showed that in-hospital mortality was statistically significantly associated with male sex (p = 0.040), ISS (p = 0.005) and mydriasis (p = 0.012). For predicting mortality, area under the curve (AUC) was calculated: for ISS 0.76 (95 % confidence interval, CI; 0.63-0.90; p < 0.001) and for initial Glasgow Coma Scale (GCS) 0.64 (95 % CI, 0.49-0.80, p = 0.079). In the multivariate analysis, ISS (odds ratio, OR; 1.19, 95 % CI, 1.06-1.35; p = 0.004) and mydriasis (OR, 5.73; 95 % CI, 1.06-30.88; p = 0.042) were identified as independent risk factors for in-hospital mortality. The AUC for the regression model was 0.83 (95 % CI, 0.71-0.94; p < 0.001). In prehospital intubated patients with isolated severe TBI only ISS and mydriasis were found to be independent predictors of in-hospital mortality.