326 Identification of Patients With Traumatic Intracranial Hemorrhage Who Are at Low Risk for Deterioration or Neurosurgical Intervention

Abstract

Current standard of care for patients with traumatic intracranial hemorrhage includes neurosurgical consultation and/or transfer to a trauma center with neurosurgical backup. We hypothesize that a set of low risk criteria can be applied to such patients to identify those who may not require neurosurgical evaluation. Pilot study of a retrospective cohort at an urban academic Level I trauma center (annual volume 92,000) of consecutive emergency department (ED) patients in 2009 with traumatic intracranial hemorrhage on computed tomography due to blunt head trauma who were >15 years old and had a GCS≥13. Charts were abstracted using a standardized data form by a single emergency physician. Our principal outcome was deterioration represented by a composite of neurosurgical intervention, clinical deterioration, or worsening computed tomography scan. Data was analyzed using Fisher's exact test. During the study period, 374 patients were seen with traumatic intracranial hemorrhage, and 204 met our inclusion criteria. Patients had a mean age of 59 years (range 16 to 97, SD=23.1), with a male predominance of 57%. Among them, 14 of 204 (6.9%) patients deteriorated. Characteristics associated with deterioration were age>64 (OR 3.44, 95% CI 1.04-11.35), history of hypertension (OR 4.59, 95% CI 1.39-15.19), abnormal mental status (OR 3.35, CI 1.04-10.77), and presence of a subdural hematoma with or without midline shift (OR 12.46, CI 1.60-97.19). No patient with isolated traumatic subarachnoid hemorrhage, isolated brain contusion or isolated skull fracture (n=63) or any combination of these 3 lesions (0/88, p<.0003) deteriorated. Use of warfarin (n=25) or aspirin (n=70) was not associated with deterioration (OR 0.83, 95% CI 0.17-3.93, OR 2.02, 95% CI 0.68-6.01, respectively). Pilot data suggest that there is a low risk group of patients with traumatic intracranial hemorrhage who may not need neurosurgical consultation or transfer to a trauma center. These include individuals with traumatic subarachnoid hemorrhage, contusion, or skull fracture; those with subdural hematoma have a higher risk of deterioration and require neurosurgical evaluation. Future efforts will be to combine these univariate predictors into a scoring model based on regression, and to attempt to validate it prospectively.