Background: Minor head injuries are a common reason for emergency department visits, with an incidence of 100 to 300 injuries per 100,000 persons. Minor head injury is defined as blunt trauma resulting in a normal or minimally altered level of consciousness at presentation to the emergency department and a maximum of a 15-minute loss of consciousness, 60 minutes of posttraumatic amnesia, or both. The evaluation of minor head injury often involves head computed tomography (CT) to rule out intracranial complications. Because most CT results in these patients are normal, there are clinical prediction rules to determine if CT is necessary. However, the most widely used rules (i.e., New Orleans Criteria and Canadian CT Head Rule) require loss of consciousness for inclusion. Smits and colleagues created a prediction rule (see accompanying table) for the selective use of CT that does not require loss of consciousness.

The Study: This prospective study included patients presenting to four teaching hospitals in the Netherlands who participated in the CHIP (CT in Head Injury Patients) trial. Patients were included if they presented within 24 hours of the injury, were at least 16 years of age, had minimally altered consciousness or normal consciousness, and had at least one of several additional criteria from previously validated prediction rules. Patients were excluded if they were transferred from another hospital or had injuries or contraindications that precluded head CT. Patients were fully examined by a neurologist or neurology resident and then received head CT based on Dutch protocol. A neuroradiologist or trauma radiologist interpreted the scans, and clinical data were entered into the study database. The authors analyzed the effects of 25 variables on the primary outcome of any intracranial traumatic finding on head CT.

Results: Of the 6,936 patients who presented with head injury during the study, 3,181 met inclusion and analysis criteria. There were intracranial CT findings in 243 patients (7.6 percent), including skull fractures, acute subdural and epidural hematomas, subarachnoid hemorrhage, and hemorrhagic contusions. Seventeen patients required neurosurgical intervention. At a cutoff score with 100 percent sensitivity for neurosurgical intervention, the prediction model missed 14 patients with intracranial traumatic CT findings. Twelve of the patients were admitted for observation; the outcome was known for 10 of these patients, all of whom had a good recovery.

Conclusion: The authors conclude that the CHIP prediction rule for selective use of CT can be applied to most patients with minor head trauma, even those without loss of consciousness. Although the rule performed well in the internal validation analysis, the authors caution that it has not been externally validated in other populations.