More than half a million head injuries occur in the United States every year, but most are minor. Following a minimal head injury, initial computed tomographic (CT) scanning identifies about 15 percent of patients as having significant intracranial conditions. A small subset of these patients require emergency craniotomy, but optimal management of the remainder has not been established. Current practice is to admit patients to the hospital for neurologic observation and serial CT. Sifri and colleagues questioned the value of repeat CT in patients with normal or improving neurologic examinations following a minimal head injury.

They studied all patients 16 years and older who had been admitted to a trauma center over 32 months with intracranial injury on CT scans following a minimal head injury. The study defined minimal head injury as loss of consciousness and/or posttraumatic amnesia, with a Glasgow Coma Score of 14 or 15. The only exclusions from the study were patients with coagulopathies (including the use of anticoagulants) and patients with a history of brain injury. All patients were assessed by neurosurgeons on admission. Patients who required immediate neurosurgical intervention were removed from the study. The remaining patients were observed with formal neurologic assessments at 12 and 24 hours plus repeat CT within 24 hours of admission.

Of 1,596 patients assessed following minimal head injury, 243 (15 percent) had an intracranial injury on the initial CT scan. Of these, 16 were excluded because of a history of brain injury or coagulopathy. After excluding patients who required immediate neurosurgical intervention and patients in whom the initial CT scan was reinterpreted as normal or showing insignificant abnormalities, 202 patients were observed and scheduled for repeat CT within 24 hours.

After a 24-hour observation period, 151 patients (75 percent) had a normal or improving neurologic examination, while 51 patients (25 percent) had an abnormal or deteriorating neurologic status. In the group of patients with normal or improving neurologic status, the follow-up CT scan was worse in 22 patients (15 percent), unchanged in 79 patients (52 percent), and improved in 50 patients (33 percent). No patients in this group required neurosurgical intervention. Conversely, among the patients who had persistently abnormal or deteriorating neurologic examination, the repeat CT scan was abnormal in 35 percent, and five patients (10 percent) required neurosurgical intervention.

The authors calculate that the negative predictive value (NPV) of normal or improving neurologic examination is 100 percent with regard to the lack of need for neurosurgical intervention. The NPV of an improving or unchanged second CT scan is 99 percent. They conclude that repeat CT scans are not indicated in patients with minimal head injury who have persistently normal or improving results on neurologic examination.