Practical selection criteria for unenhanced cranial CT in patients with acute headache

We studied clinical predictors of cranial computed tomography (CT) abnormalities in patients with acute or acutely worsened headache. Data were collected from chart review of 333 consecutive patients presenting to an emergency department and who were clinically selected for cranial CT. Patients with a positive neurologic examination were at 10.7 times greater risk for a positive CT than the rest of the sample (p<1.5 – 10⁻¹⁰). Using only neurologic examination to select patients for CT would have missed 30.3% of the positive scans. The amnesia, depressed sensorium, and hypertension variables had CT yields approximating 10% or greater even in the presence of a negative neurologic examination. Together with a positive neurologic examination, these variables detected 87.9% of the patients in this sample with positive scans; their absence had a negative predictive value of 98.0%. Of the four patients with positive scans who would have been missed using this strategy, one was discharged directly from the emergency department anyway and the other three developed positive neurologic examinations within 24 hours. One died of causes unrelated to the intracranial pathology. Positive neurologic examination, hypertension, history of amnesia, or a depressed sensorium provide reasonable initial guidelines to select for CT patients with an acute headache.