Contribution of Migraine to CVD Risk Prediction

Quick Takes

  • Migraine with aura is associated with increased risk for CVD; these data suggest this information added value when combined with CV risk prediction tools.
  • Adding migraine with aura information improved discrimination of both the Reynold Risk Score model and the AHA/ACC score model.
  • A small but significant improvement in integrated discrimination improvement and continuous net reclassification improvement (NRI) was observed; however, no improvement in categorical NRI was noted.

Study Questions:

Does the addition of migraine with aura occurrence add value when added to existing cardiovascular (CV) risk prediction tools?


Data from the WHS (Women’s Health Study), a randomized, placebo-controlled trial using a 2-by-2 factorial design to evaluate aspirin and vitamin E for primary prevention of CV disease (CVD) and cancer were used for the present analysis. Women without a history of CVD, cancer, or other major condition and without missing data used in risk factor tools were included. WHS ended in 2004, and observational follow-up is ongoing. Participants in the WHS self-reported migraine with aura status and were followed for incident CVD events. Migraine with aura status was included as a covariable in the Reynolds Risk Score and the American Heart Association/American College of Cardiology (AHA/ACC) pooled cohort equation and assessed discrimination (Harrell c-index), continuous and categorical net reclassification improvement (NRI), and integrated discrimination improvement (IDI).


A total of 24,493 female WHS participants were included in the analysis, of which 569 CVD events occurred during follow-up (median 10.2 years). Among women with self-reported migraine with aura, there were 45 events, including 20 strokes, 19 myocardial infarctions (MIs), and 8 CVD-related deaths. Only 5.1% of participants reported migraine with aura; these women tended to be younger, have a family history of MI, and have a slightly higher median C-reactive protein, but were less likely to have a history of diabetes compared to women without migraine with aura. Migraine with aura was significantly associated with CVD after including covariables in the Reynolds Risk Score (hazard ratio [HR], 2.09; 95% confidence interval [CI], 1.54-2.84) and the AHA/ACC score (HR, 2.10; 95% CI, 1.55-2.85). Adding information on migraine with aura status improved discrimination of the Reynolds Risk Score model (c-index from 0.792 to 0797; p = 0.02) and the AHA/ACC score model (c-index from 0.793 to 0.798; p = 0.01). A small but statistically significant improvement was observed in the IDI and continuous NRI after adding migraine with aura status to both models. However, no significant improvements in the categorical NRI were observed.


The authors conclude that adding information on migraine aura status to commonly used CVD risk prediction algorithms enhances model fit but does not substantially improve risk stratification among women. Despite the strong association of migraine with CVD risk, the relatively low prevalence of migraine with aura compared with other CV risk factors limits its usefulness in improving risk classification at the population level.


These findings support completing a comprehensive history to evaluate a woman’s risk for CVD, including a history of migraine with aura. As the authors note, small numbers of women reported migraine with aura; therefore, further data on women with documented migraine with aura and from a diverse population are warranted.

Clinical Topics: Prevention

Keywords: Cardiovascular Diseases, C-Reactive Protein, Diabetes Mellitus, Migraine Disorders, Migraine with Aura, Myocardial Infarction, Primary Prevention, Risk Assessment, Risk Factors, Stroke, Women

< Back to Listings