Article
Prediction of post-concussive symptoms in children and adolescents with traumatic brain injury – a CENTER-TBI study analysis
Vorhersage von postkommotionellen Symptomen bei Kindern und Jugendlichen mit Schädel-Hirn-Trauma – eine Analyse der CENTER-TBI Studie
Search Medline for
Authors
Published: | June 26, 2020 |
---|
Outline
Text
Objective: Post-concussive symptoms are frequent in patients with traumatic brain injury (TBI). While various efforts have been made to develop clinical prediction models for post-concussive syndrome (PCS) in adult TBI, data on this condition in children or adolescents is scarce. We, therefore, aimed to develop a set of clinical predictors that would allow identifying pediatric or adolescent TBI patients with a high risk for PCS.
Methods: The multicenter prospective CENTER-TBI database was screened and only patients < 25 years with available Rivermead Post-Concussion Questionnaire (RPQ) at 6-months were included. PCS was dichotomized and defined as having at least three of the sixteen symptoms included in the RPQ. Potential predictors were selected based on existing literature and regression analysis with the least absolute shrinkage and selection operator (LASSO) method was performed to select a multivariate model with the most important determinants predicting PCS. Model performance was assessed using bootstrap validation.
Results: 424 eligible young TBI patients were identified (median age 19 (IQR 16-22) years, 70% males). Age, gender, history of psychiatric illness, GCS score at admission, abnormality on CT imaging, post-traumatic amnesia, loss of consciousness, and ICU admission were found to be the strongest predictors for PCS after LASSO model selection. This model displayed reasonable discrimination (area under the receiver operating curve of 0.70) but explained only 16% of the variance in outcome after bootstrap validation.
Conclusion: A set of 8 eight clinical factors easily obtainable in the emergency room is predictive for PCS in children and adolescents. Using these predictors might help to identify young patients that would benefit from early follow-up appointments with e.g. neurocognitive testing. However, these factors only explain a small part of the variation in outcome after TBI and additional variables such as biomarkers might be needed to improve outcome prediction.