Abstract 47: Using Ultrasound and Inflammation to Improve Prediction of Ischemic Stroke: A Secondary Analysis of MESA

Abstract

Background: The Multi-Ethnic Study of Atherosclerosis (MESA) is a large prospective epidemiologic study of the clinical factors that can predict transition from asymptomatic to symptomatic cardiovascular disease. Although prior studies have looked at ischemic stroke, they have not systematically examined the relationship between baseline ultrasound and inflammation measurements and subsequent primary stroke risk. Methods: The primary outcome is incident ischemic stroke during follow-up. The predictors are 9 ultrasound-derived measurements and 5 serum measurements related to inflammation. We fit Cox models to ischemic stroke and adjusted for patient age, hypertension, diabetes, total cholesterol, and smoking. Using DeLong’s method, we compared the AUC of the baseline adjusted model to the AUC of the model with predictor variables that were significant in the Cox models, to determine if they improved stroke prediction. Results: We included 6,095 patients with an average age of 61.9 years. The primary outcome of ischemic stroke was seen in 107 patients (1.8%) and the mean follow-up time was 7.7 years. In the Cox models, we found that small artery elasticity (SAE), carotid distensibility (CD), carotid stenosis (CS), and interleukin-6 (IL6) were associated with incident stroke. The AUC of the baseline model to predict stroke, which included patient age, hypertension, diabetes, total cholesterol, and smoking, was 0.745. When we added tertiles of SAE, CD, IL6 and categories of CS, the AUC improved to 0.765 (p=0.021 for difference). Conclusions: In a multiethnic cohort of patients without CVD at baseline, we found several ultrasound measurements and a serum marker of inflammation which predicted the occurrence of a primary ischemic stroke. Adding these basic ultrasound and serum measurements significantly improved the prediction of stroke, which could have implications for primary prevention efforts.