The comorbidity of migraine and incident cardiovascular disease (CVD) has motivated a search to understand potential shared pathophysiologic mechanisms. These explorations have focused on two of the most important classes of CVD risk factors, plasma lipid levels and blood pressure (BP). Of the two classes, associations between migraine and lipids have been more consistent, implicating elevated LDL cholesterol (LDLC) and triglycerides but not HDL cholesterol in migraine. Within this context, a recent randomized trial among migraineurs found that the combination of simvastatin, an LDLC lowering agent, and vitamin D significantly reduced migraine frequency compared to placebo. In contrast, observational studies of BP (systolic BP, diastolic BP, and pulse pressure) and migraine, which are susceptible to bias, have shown positive, negative, or null associations. Despite this mixed evidence, BP-lowering drugs reduce migraine frequency while genetic analyses highlight the vascular system in migraine pathophysiology, both suggesting connections to BP. These results raise two, non-exclusive, possibilities for potential links between the risk factors and migraine: there may be a causal (or reverse causal) relationships and/or there may be shared underlying biology. We will implement recent analytic approaches leveraging genetics to evaluate the extent to which each of these two scenarios may be relevant. The first analytic approach termed ?Mendelian Randomization?, exploits the randomized and irreversible meiotic assortment of alleles influencing lipid levels or BP to assess whether lifelong exposure to a genetically determined risk factor may be associated with migraine, thereby implying causality. Using individual level information for migraine, lipid levels, BP, and genome-wide genotype data from the Women's Genome Health Study (5,122 migraineurs/23,294 total) and the UK Biobank (14,392 migraineurs/~500,000 total), we will perform MR, including tests of reverse causality, to compare observational effects with genetically inferred causal effects, which would be further characterized by mediation analysis. These analyses will be complemented by MR analysis using summary statistics from extremely well-powered, genome-wide association studies of migraine (>56,000 migraineurs), lipid levels (up to 188,577 individuals), and BP (up to 342,415 individuals). In the second analytic approach, we will estimate the extent of shared genetic influence, or ?shared heritability?, between each risk factor and migraine in these same data sets while also identifying specific genetic variants with shared associations. Moreover, we will use computational approaches to infer which tissue types and/or biological pathways may be most relevant to the shared genetic effects. Due to the robustness of the analytic approaches and the statistical power of available data, the proposed research is expected to reach definitive conclusions about relationships between migraine and lipids or BP, and is likely to advance understanding of migraine pathophysiology if not also to suggest novel therapeutic strategies for migraine.
There is compelling evidence that modifying two classes of vascular risk factors, plasma lipid levels and blood pressure, even with existing drugs, can be part of a therapeutic strategy for migraine. However, the mechanistic basis of the relationship between these risk factors and migraine is unknown. The proposed research is relevant to public health for its potential to reveal the mechanistic basis of the connections between these risk factors and migraine, and therefore inform on their potential as targets for migraine therapy.
