Guidelines for minor head trauma CT fall short

No one version stands out from the pack of European and international guidelines for the use of CT in minor head trauma because of trade-offs between sensitivity and specificity. Dutch researchers determined that the sensitivity of their own guidelines for neurosurgical interventions was unacceptably low.

Dr. Marion Smits and colleagues at Erasmus Medical Center conducted a prospective multicenter study in four university hospitals in the Netherlands of consecutive adult patients with minor head injury.

The researchers evaluated head CT scans of 3181 patients who presented with a Glasgow Coma Score of 13 to 14 or a GCS of 15 and at least one risk factor: loss of consciousness, amnesia, headache, vomiting, short-term memory deficit, seizure, neurological deficit, intoxication, supraclavicular injury, or coagulopathy.

The primary outcome measure was any neurocranial traumatic finding on CT. Secondary outcome measure was any traumatic finding on CT that led to neurosurgical intervention.

Traumatic findings appeared in 312 patients (9.8%), while neurosurgical intervention was performed in 17 patients (0.5%).

The researchers assessed several sets of guidelines:

• Dutch national

• European Federation of Neurological Societies

• Neurotraumatology Committee of the World Federation of Neurosurgical Societies (WFNS)

• National Institute for Clinical Excellence, U.K.

• Scandinavian

• Scottish Intercollegiate Guidelines Network

Only the European Federation guidelines reached a sensitivity of 100% for both neurosurgical intervention and traumatic lesions on CT, but at the cost of 0% specificity.

The Scottish guidelines for CT findings had similar high sensitivity (99%) and low specificity (2.4%), as did the World Federation guidelines (98.4% and 2.9%, respectively).

The best specificities for traumatic lesions and neurosurgical interventions were reached with the U.K criteria (46.2% and 43.6%, respectively), but at the cost of relatively low sensitivities (82.1% and 94.1%, respectively), Smits said.

The Dutch guidelines for neurosurgical interventions had the lowest sensitivity (76%), with 23% specificity.

Researchers are now investigating the cost-effectiveness associated with these guidelines.