Abstract
Antidepressants are among the most-prescribed drugs worldwide, and selective serotonin reuptake inhibitors (SSRIs) are among the most prescribed antidepressants, most commonly used for major depression. We sought to increase our understanding of the biological relationships between SSRI use and a range of psychiatric traits by conducting a genome-wide association study (GWAS) in two large datasets, the UK Biobank (UKB) and the US Million Veteran Program (MVP). GWAS was conducted across 22 autosomes and the X chromosome in 777,952 individuals of European ancestry (191,800 SSRI users, 586,152 control individuals) and 112,526 individuals of African ancestry (53,499 SSRI users, 59,027 control individuals). We identified 40 genome-wide significant (GWS) loci, including two on the X chromosome. Using linkage disequilibrium score regression, we detected strong genetic correlations between MVP and the independent UKB cohort for specific SSRIs (fluoxetine rg = 0.82, citalopram rG = 0.89) as well as with headaches (rG = 0.80), major depressive disorder (MDD; rG = 0.77), and spondylosis (rG = 0.84), suggesting stability in trait definition across cohorts. To evaluate differences in genomic variance captured by SSRI use vs. MDD, we performed a comparative rG analysis and found significant differences, most notably for educational attainment (SSRI rG = -0.38, MDD rG = -0.26), cognitive performance (SSRI rG = -0.31, MDD rG = -0.15), and depression (SSRI rG = 0.80, MDD rG = 0.97). In MVP, SSRI use showed greater locus discovery than MDD (28 vs. 17 loci); comparison to a prior GWAS of anxiety symptoms identified only five loci. SSRI use is likely a partial proxy for MDD, while also reflecting distinct features relevant to related disorders such as anxiety.