Associations of sleep, sedentary behavior, and physical activity with incident migraine: a prospective cohort study

Abstract Background Migraine, a common primary headache disorder, has implicated daily behaviors like sleep, physical activity (PA), and sedentary behavior (SB), yet the findings are inconsistent or remain unclear. This prospective cohort study (2006–2022) aimed to examine the independent and joint associations of sleep, PA, and SB with migraine risk in a general adult population. Methods Data of 484,588 participants without baseline migraine from the UK Biobank were included. Multiple Cox regressions, restricted cubic spline models, and isotemporal substitution model (ISM) were used to assess associations between sleep, PA, SB and incident migraine. A nonlinear association model was used to analyze the correlation between each behavior and magnetic resonance imaging (MRI) brain volume and visualized using the “ENIGMA” tool. Results During a median follow-up of 13.63 years, 4,380 participants developed migraine. Sleep duration, PA, and SB exhibited distinct nonlinear associations with migraine risk. The risk of migraine was 40% lower in the healthy sleep score group (adjusted hazard ratio (aHR)[95% confidence interval (CI)]: 0.60[0.53, 0.69], p < 0.001) and 23% higher in the high SB group (aHR[95%CI]: 1.23[1.11, 1.37], p < 0.001) compared to the corresponding reference group. Upon examining the joint effects between sleep patterns and SB, individuals with low SB and poor sleep score showed the higher risk of migraine (aHR[95%CI]: 1.98[1.30, 3.02], p = 0.003). Conclusion In this large prospective cohort, healthier sleep patterns and lower levels of SB were associated with a reduced risk of incident migraine. PA exhibited a complex nonlinear association that did not remain significant in categorical analyses after full adjustment. Notably, sleep duration and SB were associated with structural variations in brain regions implicated in migraine. These findings provide a large-scale epidemiological map linking sleep and SB to migraine risk, and generate the hypothesis that these behaviors are modifiable targets for future preventive intervention trials.