Vascular Risk Prediction in Migraine: A Multimodality Risk Score Based on Artificial Intelligence-electrocardiogram Output, Echocardiography, and Detailed Migraine Characteristics
Keiko Ihara1, Nan Zhang2, Ping-Hao Yang1, Chieh-Ju Chao3, Francisco Lopez-Jimenez3, Kathryn E Mangold3, Zachi Attia3, Paul A Friedman3, Peter A Noseworthy3, Chia-Chun Chiang1
1Department of Neurology, 2Department of Quantitative Health Sciences, 3Department of Cardiovascular Medicine, Mayo Clinic
Objective:
To develop a multimodality model to predict vascular risks in patients with migraine based on comprehensive clinical information, including artificial intelligence (AI)-electrocardiogram (ECG) output, echocardiogram parameters, and migraine characteristics.
Background:
Patients with migraine have higher risks for stroke and other adverse vascular events. We have conducted prior analyses on AI-ECG algorithms, echocardiogram parameters, and detailed migraine characteristics separately to predict vascular events. However, the relative importance and combined effects of each modality have not been established.
Design/Methods:
We conducted retrospective observational study on patients with at least one 12-lead ECG, transthoracic echocardiogram and a headache specialist visit at any Mayo Clinic site within 3 years. The outcome is a composite endpoint of acute ischemic stroke, acute myocardial infarction, venous thromboembolism, and all-cause mortality. We selected 20 risk parameters, including (a) demographics, (b) comorbidities, (c) AI-ECG output, (d) echocardiogram, (e) migraine characteristics found to be predictive of vascular events. Each variable was assigned a different risk score according to the hazard ratio (HR) and beta coefficient from prior Cox regression analysis. Additionally, a multivariate Cox regression was conducted to assess the relative risk of each parameter category.
Results:
We included 297 patients with migraine, of whom 247 were female and 109 had migraine with aura. The mean age was 39.0±13.3 and follow-up period was 68.9±60.9 months. Thirty-five developed the composite endpoint (11.8%). Those with a risk score ≥15 had significantly shorter vascular event-free survival (HR [95% CI], 3.85 [1.90–7.81]; p<0.001) compared to those scored <15. Furthermore, comorbidities (HR, 1.17 [1.06–1.30]; p=0.003) and migraine characteristics (HR, 1.34 [1.05–1.71]; p=0.018) were significantly predictive of the composite endpoint, whereas other categories were not.
Conclusions:
This study supports the use of AI-ECG output, echocardiogram and migraine characteristics for vascular risk prediction, and highlights the importance of detailed migraine information for risk stratification.
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.