Exploring the genetic overlap between attention-deficit/hyperactivity disorder (ADHD) and migraine

Abstract Background Attention-deficit/hyperactivity disorder (ADHD) and migraine are prevalent neurodevelopmental and neurological conditions, respectively, that contribute to individual disability and social burden. The biological mechanisms linking these disorders remain poorly understood. Methods We aimed to investigate their shared genetic architecture by integrating genomic data with a cross-trait analysis using the largest genome-wide association studies (GWAS) for ADHD and migraine to date. Variants were classified into concordant and discordant, according their direction of effect, and were followed-up with functional and colocalization analyses. Then, polygenic risk scores (PRSs) analyses were undertaken aiming to dissect their clinical heterogeneity, using in-house ADHD and migraine samples. Results We identified 961 pleiotropic SNPs across 29 loci, including 9 loci not previously related to either condition. Concordant variants were enriched in immune-related pathways, brain morphology, and autoimmune traits, while discordant variants mapped to genes associated with psychiatric, cardiovascular, and behavioural traits. Colocalization analysis showed more pleiotropic concordant loci than discordant loci (62% vs 17%). Individuals with higher PRS for migraine showed increased odds for ADHD. Also, migraine PRSs were positively associated to childhood headaches in individuals with ADHD. This association was stronger when restricting the PRS to concordant migraine-specific and pleiotropic variants, despite including only one quarter of them. Furthermore, ADHD individuals reporting childhood headaches also showed higher levels of anxiety, depression, neuroticism, and reduced cognitive performance in adulthood. Conclusions These findings suggest shared genetic mechanisms between ADHD and migraine, particularly involving neuroimmune and neurodevelopmental pathways, and support the utility of pleiotropy-informed PRSs models for understanding comorbid traits.