Kerstin de Wit: Evaluating a selective CT head scanning approach in anticoagulated head-injured emergency department patients


The population who is prescribed anticoagulation medication is growing. Current guidelines on neuroimaging recommend head computed tomography (CT) for all anticoagulated patients with head injuries, however these guidelines were based on warfarin. Today, more patients are prescribed direct oral anticoagulants (DOACS) than warfarin. DOACs have a lower risk of intracranial bleeding than warfarin. We aim to expand the applicability of head injury clinical decision rules to anticoagulated patients, to ensure that evidence-based neuroimaging decisions can be made for this rapidly growing patient population. 


This is a prospective international cohort study conducted in 23 hospitals. We will enroll anticoagulated ED patients presenting within 48 hours of a blunt head injury. Patients must have a GCS of 13-15, or else be at their usual baseline GCS, and be current anticoagulant users. Exclusion criteria are age <16, penetrating skull injury, transfer from another ED, prior enrolment in the study, patients who leave the ED prior to completion of their medical assessment and patients who live outside of the hospital catchment area.

The physician will document the patient clinical findings on a standard form prior to CT scanning and demographic data will be captured by structured electronic chart review. 

The primary outcome is clinically relevant traumatic brain injury as determined by a blinded independent adjudication committee. The secondary outcome is delayed clinically relevant traumatic brain injury diagnosed within 30 days.

We will report the sensitivity, specificity, efficiency and false negative rate of the Canadian CT Head Rule. We will also assess the classification performance of the New Orleans, NEXUS II, CHIP and NICE decision rules. The primary outcome is sensitivity and specificity for clinically relevant traumatic brain injury. Secondary outcomes classification performance of the rules in the subgroup of patients taking both anticoagulant and antiplatelet medications. Secondary outcomes also include the performance of each rule in diagnosing delayed traumatic brain injury. If no rule has a sensitivity of >98% and a specificity of >25% for clinically relevant traumatic brain injury, we will derive a new rule for this patient group using multivariable logistic regression modeling.

Impact on Emergency Medicine

A safe and efficient clinical decision rule could reduce the emergency department length of stay, radiation exposure for patients and health system costs.