Genome-Wide Association Study of Generalized Anxiety Disorder Symptoms and Self-Reported Anxiety Disorder: An Investigation of Unique and Shared Genomic Contributions

Abstract

Generalised Anxiety Disorder (GAD) is a pervasive condition characterised by states of consistently high anxiety across multiple situations. GAD is highly comorbid with other Anxiety Disorders (AD’s) with approximately 60% of GAD patients being additionally diagnosed with one of these disorders (Wittchen et al. 2011). In addition to phenotypic overlap, GAD shows significant genetic overlap with other AD’s (Waszczuk et al. 2014). There is evidence that traits central to GAD (such as heightened anxiety sensitivity) are associated with specific genetic variants (Davies et al. 2015). A meta-analysis of genome-wide association studies of AD’s identified loci associated with anxiety case status, and with a common anxiety factor score (Otowa et al. 2016).Thus, there is evidence that core components of GAD symptom spectrum are associated with variation in specific genetic loci, and that AD’s share common genetic factors. To date, no studies have investigated GAD-specific genome-wide associations, nor the overlap between these and other AD’s.The current study aims to analyse the association between GAD and genome-wide single nucleotide polymorphisms (SNP’s), and investigate the extent to which these associations are unique to GAD, or are shared across AD’s.Analyses were undertaken in approximately 138,811 individuals from the UK Biobank. Three approaches to defining anxiety phenotypes were adopted. (1) Adjusted raw scores on the Generalised Anxiety Disorder-7 (GAD-7) were used for all participants, resulting in a continuous quantitative GAD-7 phenotype, (2) GAD case status was established using responses to items from the composite international diagnostic interview – short form (CIDI - SF), and (3) AD case status was established through self-report of any life time diagnosis of any AD.Associations between adjusted GAD-7 total scores and genome wide SNP’s were analysed using linear regression. Associations between GAD and AD case status and genome-wide SNP’s were analysed using logistic regression. Gene pathway analyses were conducted for significant hits, and SNP heritability was estimated for all phenotypes. The correlation between SNP heritability of GAD and non-specific AD diagnoses is investigated.