Improved polygenic risk prediction in migraine-first patients.

Background: Recent meta-analyses estimated 14.6% and 11.2% SNP-based heritability of migraine, compared to twin-heritability estimates of 30-60%. This study aimed to investigate heritability estimates in "migraine-first" individuals, patients for whom G43 (migraine with or without aura) was their first medical diagnosis in their lifetime.

Findings: Using data from the UK Biobank (N = 199,929), genome-wide association studies (GWAS) were conducted on 6,139 migraine-first patients and 193,790 healthy controls. SNP-based heritability was estimated using SumHer, yielding 19.37% (± 0.019) for all SNPs and 21.31% (± 0.019) for HapMap3 variants, substantially surpassing previous estimates. Key risk loci included PRDM16, FHL5, ASTN2, STAT6/LRP1, and SLC24A3, and pathway analyses highlighted retinol metabolism and steroid hormone biosynthesis as important pathways in these patients.

Conclusions: The findings underscore that excluding comorbidities at onset time can enhance heritability estimates and genetic signal detection, significantly reducing the extent of "missing heritability" in migraine.