A new clinical score for cranial CT in ED non-trauma patients: Definition and first validation.

ABSTRACT

INTRODUCTION: Well recognized guidelines are available for the use of cranial computed tomography (CCT) in traumatic patients, while no definitely accepted standards exists to for CCT in patients without history of head injury. The aim of this study is to propose an easy clinical score to stratify the need of CCT in emergency department (ED) patients with suspect non-traumatic intracranial pathology. METHODS: We retrospectively evaluated patients presenting to the ED for neurological deficit, postural instability, acute headache, altered mental status, seizures, confusion, dizziness, vertigo, syncope, and pre-syncope. We build a score for positive CCT prediction by using a logistic regression model on clinical factors significant at univariate analysis. The score was validated on a population of prospectively observed patients. RESULTS: We reviewed clinical data of 1156 patients; positivity of CCT was 15.2%. Persistent neurological deficit, new onset acute headache, seizures and/or altered state of consciousness, and transient neurological disorders were independent predictors of positive CCT. We observed 508 patients in a validation prospective cohort; CCT was positive in 11.3%. Our score performed well in validation population with a ROC AUC of 0.787 (CI 95% 0.748-0.822). Avoiding CT in score 0 patients would have saved 82 (16.2%) exams. No patients with score 0 had a positive CCT findings; score sensitivity was 100.0 (CI 95% 93.7-100.0). CONCLUSIONS: A score for risk stratification of patients with suspect of intra-cranial pathology could reduce CT request in ED, avoiding a significant number of CCT while minimizing the risk of missing positive results.