Plasma proteomics identifies proteins and pathways associated with incident migraine in 50,668 adults

Migraine is a prevalent neurological disorder, but the systemic molecular alterations that precede its onset and their biological relevance remain incompletely understood. We conducted a prospective plasma proteomics analysis among 50,668 migraine-free participants from the UK Biobank. Baseline levels of 2,923 plasma proteins were quantified using the Olink Explore platform. Multivariable Cox proportional hazards models were applied to assess associations between protein levels and incident migraine during follow-up, with adjustment for demographic, lifestyle, and clinical covariates. False discovery rate (FDR) correction was used to account for multiple testing. Trajectory analyses were performed to characterize temporal protein patterns prior to diagnosis. Functional enrichment, protein–protein interaction, and transcription factor analyses were conducted to explore biological pathways. Single-cell transcriptomic data were integrated to assess cell-type specificity. Genetic analyses, including polygenic risk score associations, Mendelian randomization, and Bayesian colocalization, were used to evaluate potential causal relationships. Associations between migraine-related proteins and global brain structural measures derived from magnetic resonance imaging were examined. Machine-learning models were developed to evaluate the predictive performance of plasma proteomic profiles for future migraine risk. After FDR correction, 126 proteins were significantly associated with incident migraine, of which 124 were retained for downstream analyses. Distinct protein trajectories were observed several years before clinical diagnosis. Enrichment analyses consistently implicated immune regulation and neuro-immune signaling pathways. Genetic analyses supported potential causal roles for a subset of proteins, and several migraine-associated proteins were linked to global brain structural measures. Proteomics-based models achieved modest but reproducible discrimination of future migraine risk. These findings demonstrate that migraine is preceded by long-term systemic proteomic alterations, highlight immune-related molecular pathways and genetically supported proteins potentially involved in migraine pathogenesis, and provide a foundation for future mechanistic studies and early risk stratification strategies. Not applicable.