|
Abstract
|
BACKGROUND: The majority of clinical decision rules for prediction of intracranial injury in patients with mild traumatic brain injury (TBI) were developed from high-income countries. The application of these rules in low or middle-income countries, where the primary mechanism of injury was traffic accidents, is questionable.
METHODS: We developed two practical decision rules from a secondary analysis of a multicenter, prospective cohort of 1,164 patients with mild TBI who visited the emergency departments from 2013 to 2016. The clinical endpoints were the presence of any intracranial injury on CT scans and the requirement of neurosurgical interventions within seven days of onset.
RESULTS: Thirteen predictors were included in both models, which were age ≥60 years, dangerous mechanism of injury, diffuse headache, vomiting >2 episodes, loss of consciousness, posttraumatic amnesia, posttraumatic seizure, history of anticoagulant use, presence of neurological deficits, significant wound at the scalp, signs of skull base fracture, palpable stepping at the skull, and GCS <15 at 2 hours. For the model-based score, the area under the receiver operating characteristic curve (AuROC) was 0.85 (95%CI 0.82, 0.87) for positive CT results and 0.87 (95%CI 0.83, 0.91) for requirement of neurosurgical intervention. For the clinical-based score, the AuROC for positive CT results and requirement of neurosurgical intervention was 0.82 (95%CI 0.79, 0.85) and 0.84 (95%CI 0.80, 0.88), respectively.
CONCLUSIONS: The models delivered good calibration and excellent discrimination both in the development and internal validation cohort. These rules can be used as assisting tools in risk stratification of patients with mild TBI to be sent for CT scans or admitted for clinical observation.
Language: en |