Epidemiology and early predictive factors of mortality and outcome in children with traumatic severe brain injury: experience of a French pediatric trauma center.

摘要

OBJECTIVE: To describe the results of an integrated pre- and in-hospital approach to critical care in a large population of children with severe traumatic brain injury and to identify the early predictors of their outcome. DESIGN: A 9-yr retrospective review of the data of a trauma data bank. SETTING: Level III pediatric trauma center. PATIENTS: All children (1 month to 15 yrs) with severe traumatic brain injury (Glasgow Coma Scale /=6 months after discharge. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Univariate and further multivariate analyses were performed to determine independent predictive factors of death and outcome at discharge and 6 months later. The Glasgow Outcome Scale was used to evaluate outcome; a poor outcome referred to Glasgow Outcome Scale >/=3. Receiver operating characteristic curves were drawn to determine the threshold values of predictors of death and outcome. Analysis concerned 585 children (67% male and 33% female). Mean age was 7 +/- 5 yrs. Predominant mechanisms of injury were road traffic accidents and falls. Mean values for Glasgow Coma Scale, Pediatric Trauma Score, and Injury Severity Score were 6 (3-8), 3 (-4,10), and 28 (4-75), respectively. Mortality rate was 22%; Glasgow Outcome Scale was <3 in 53% of the cases at discharge and 60% at 6 months. Multivariate analysis identified Glasgow Coma Scale, Injury Severity Score, and hypotension on arrival as independent predictors of death and poor outcome at discharge and at 6 months. Threshold values for death were 28 for Injury Severity Score and 5 for Glasgow Coma Scale. The same values were found for poor outcome, except for outcome at 6 months where threshold value for the Glasgow Coma Scale was 6. CONCLUSIONS: Initial hypotension, Glasgow Coma Scale, and Injury Severity Score are independent predictors of outcome in children with traumatic brain injury. Threshold values can be calculated for predicting poor outcome. These variables can be easily and detected early in this population and used for quality assessment.