Clinical Report: Stroke Risk Assessment in Patients with Acute Persistent Vertigo
Overview
This study developed and internally validated a multivariable predictive model for assessing stroke risk in patients with acute persistent vertigo. The model demonstrated superior performance compared to existing scoring systems.
Background
Acute persistent vertigo can indicate a stroke, yet effective risk assessment remains challenging. Timely identification of strokes in patients presenting with vertigo is crucial for optimizing treatment and improving outcomes. Current scoring systems often lack the necessary sensitivity and specificity for this patient population.
Data Highlights
Variable
Odds Ratio (OR)
Age
1.050 per year
Smoking
2.147
Hypertension
5.452
Hyperlipidemia
2.621
Diabetes Mellitus
3.918
Coronary Heart Disease
4.361
Atrial Fibrillation
10.376
CNS Score
1.679 per point
Nausea/Vomiting
2.020
Tinnitus
0.382
Key Findings
The multivariable model showed excellent discrimination with an AUC of 0.902.
Stroke event rates were 2.6%, 10.5%, and 58.7% in low-, intermediate-, and high-risk groups, respectively.
Older age, smoking, hypertension, and diabetes were significant predictors of stroke risk.
The model outperformed existing scoring systems such as ABCD2 and Triage-Plus.
Calibration metrics indicated good model performance with a Brier score of 0.096.
Clinical Implications
The developed predictive model may assist in stratifying stroke risk in patients with acute persistent vertigo.
Conclusion
The multivariable prediction model offers a method for stroke risk assessment in acute persistent vertigo.