Development of a clinical risk score to risk stratify for a serious cause of vertigo: A prospective cohort study

Abstract

Objectives: Identify highrisk clinical characteristics for a serious cause of vertigo in patients presenting to the emergency department. Design: Multicentre prospective cohort study over 3 years. Setting: Three university affiliated tertiary care emergency departments. Participants: Patients presenting with vertigo, dizziness or imbalance. A total of 2078 of 2618 potentially eligible patients (79.4%) were enrolled (mean age 77.1 years; 59% women). Main outcome measurements: An adjudicated serious diagnosis defined as stroke, transient ischemic attack, vertebral artery dissection or brain tumour. Results: Serious events occurred in 111 (5.3%) patients. We used logistic regression to create a 7 item prediction model: male, age over 65, hypertension, diabetes, motor/sensory deficits, cerebellar signs/symptoms and benign paroxysmal positional vertigo diagnosis (C statistic 0.96, 95% confidence interval [CI] 0.92 0.98). The risk of a serious diagnosis ranged from 0% for a score of <5, 2.1% for a score of 5-8, and 41% for a score >8. Sensitivity for a serious diagnosis was 100% (95% CI, 97.1 100%) and specificity 72.1% (95% CI, 70.1 74%) for a score <5. Conclusions: The Sudbury Vertigo Risk Score identifies the risk of a serious diagnosis as a cause of a patient's vertigo and can assist physicians in guiding further investigation, consultation and treatment decisions, improving resource utilization and reducing missed diagnoses.