Outcome of Patients With Mild Head Injury and Presence of an Acute Traumatic Abnormality on CT Scan of Head

Overview

Background: Patients with mild blunt traumatic brain injury (TBI) are frequently transferred to Level 1 trauma centers (L1TC) if they have any positive finding of any acute intracranial injury identified on a CT scan of the head. The hypothesis for the study is that patients with such injuries and minor changes on the Head CT scan can be safely managed at community hospitals (CH). Methods: Patients with blunt, mild TBI (defined as a GCS 13-15 at presentation) presenting to CH, L1TC, and transferred from CH to L1TC between March, 2012 and February, 2014 were included. Minor changes on head CT were defined as: 1) epidural hematoma\<2mm; 2) subarachnoid hemorrhage\<2mm; 3) subdural hematoma\<4mm; 4) intraparenchymal hemorrhage\<5mm; 5) minor pneumocephalus; or 6) linear or minimally depressed skull fracture. TBI-specific interventions were defined as intracranial pressure monitor placement, administration of hyperosmolar therapy, or neurosurgical operation. Three groups of patients were compared: 1) those receiving treatment at CH, 2) those transferred from CH to L1TC, and 3) those presenting directly to L1TC. The primary endpoint was the need for TBI-specific intervention and secondary outcome was death of any patient.

Methods The trauma registries at all participating centers were searched for patients who developed mild TBIs following blunt trauma and were directly admitted either to L1TC, or CH, or transferred from CH to L1TC. Patients with Glasgow Coma Scale (GCS) equal to or greater than 13 and a positive head CT scan for minor injuries were included in the study. Minor CT findings were defined as: 1) an epidural hematoma less than 2 mm thick, 2) a subarachnoid hemorrhage measuring less than 2 mm, 3) a subdural hematoma less than 4 mm thick, 4) an intraparenchymal hemorrhage measuring less than 5 mm, 5) minor pneumocephalus, or 6) linear or minimally depressed skull fracture. Patients with multiple findings were also included so long as the above criteria were met. Patients were also included patients if they were taking aspirin or if they were intoxicated with alcohol as long as their GCS could still be assessed to be between 13-15. Patients with more severe CT scan findings were excluded. Patients were also excluded if they were younger than 18 years of age, presented with open skull fractures, were intubated or hemodynamically unstable upon presentation, or had prior history of bleeding diathesis. Finally, patients with injuries in other areas of the body with an abbreviated injury score (AIS) \> 2 were excluded. After obtaining approval by the Institutional Review Board, data of interest was retrospectively collected from one LITC and four CH. This was done by using the trauma registries and reviewing individual medical charts. Collected data included baseline demographics (e.g. age and gender), variables related to the blunt trauma (e.g. mechanism of injury, injury severity score \[ISS\], and AIS scores), baseline comorbidities, vital signs and GCS on arrival to the emergency department, CT scan findings and whether a repeat CT scan of the head was performed, the administration of blood products, hospital and intensive care unit (ICU) length of stay, as well as in-hospital complication and mortality rates. Three groups of patients were compared: 1. those who were admitted and received definitive treatment at one of the four CH 2. those who initially presented at CH but were subsequently transferred to L1TC and 3. those who presented directly to the L1TC. The primary endpoint of the study was the need for TBI-specific interventions in these 3 groups. TBI-specific intervention was defined as a neurosurgical operation, insertion of an intracranial pressure (ICP) monitor, or administration of hyperosmolar therapy. The secondary endpoint was mortality. Statistical analysis was performed using the STATA software (version 13.1). Numerical variables are reported as medians with interquartile ranges (25th to 75th percentile), and categorical ones as frequencies and percentages. The Kruskal Wallis non-parametric test was used to compare the numerical variables and the chi-square or Fisher's exact test to compare the categorical variables as appropriate. The multivariable logistic regression analyses to identify independent predictors of TBI-specific interventions or independent risk factors for mortality and overall morbidity could not be performed, given the rarity of these events in the patient population. p-value of less than 0.05 was defined as the level of statistical significance.