51: POLYGENIC RISK FOR ADHD IS ASSOCIATED WITH READING AND SPELLING RELATED TRAITS BEYOND PLEIOTROPIC EFFECTS DUE TO EDUCATIONAL ATTAINMENT

Abstract

Background Language impairments often co-occur with Attention Deficit and Hyperactivity Disorder (ADHD) symptoms. Twin studies suggest that this comorbidity is at least partially due to shared genetic factors. Here, we aim to study the association between polygenic ADHD risk and language abilities in the general population. Using multivariable regression analysis, we also investigate whether this link persists conditional on the polygenic effects for Educational Attainment (EA), a factor that is known to be correlated with both ADHD and Language-Related Abilities (LRAs). Methods Thirteen LRAs capturing comprehension, reasoning, reading, spelling, phonological awareness/memory and verbal IQ were studied in 7 to 13 year-old children from a UK birth cohort (ALSPAC; N≤5,919; SNP-h2>0.25). ADHD summary statistics were obtained from the Psychiatric Genomics Consortium (PGC; cases:4,163; controls:12,040), the Lundbeck Foundation Initiative for Integrative Psychiatric Research (iPSYCH; cases:14,584; controls:22,492) and a combined sample thereof (PGC+iPSYCH; cases:20,183; controls:35,191). EA summary statistics excluding ALSPAC participants were obtained through the Social Science Genetic Association Consortium (N=326,041). Genetic overlap between ADHD and LRAs was assessed using polygenic scoring (PGS). Conditional associations of genetically-predicted ADHD and EA with LRAs were estimated using Mendelian Randomisation techniques. However, the current data allows us to interpret findings in terms of shared genetic liability only. Effect estimates for spelling, reading and overall LRAs were combined using random-effects meta-regression, weighted by phenotypic correlations. Results Across all three ADHD samples, we consistently observed an inverse association between ADHD risk increasing alleles and both reading (e.g. PGC+iPSYCH; P=4×10–19; β=-0.11; max OLS-R2 =1.5%) and spelling (PGC+iPSYCH; P=4×10–18; β=-0.11; max OLS-R2 =1.2%) performance, but not overall LRAs, using PGS. Conditional analyses on EA confirmed inverse associations between ADHD polygenic risk and reading using both robust (15 SNPs; P-threshold Discussion Polygenic risk for ADHD shares genetic liability with reading- and spelling-related traits due to both disorder specific genetic effects but also pleiotropic effects with EA.