TRANSLATING DISCOVERIES IN ADHD GENOMICS TO THE CLINIC

Background A recent genome-wide association analysis has identified specific susceptibility variants and a substantial polygenic component for Attention-Deficit/Hyperactivity Disorder (ADHD). Given that emerging risk variants are unlikely to track precisely with conventional diagnoses, whether these discoveries will have utility in clinical settings for improved diagnostics and risk stratification requires further study. Our overall goal is to establish the relevance of emerging psychiatric genetics findings to youth clinical samples. In the current study, we aimed to 1) confirm the convergent validity of ADHD polygenic risk with ADHD-related phenotypes in youth presenting for neuropsychiatric evaluation; and 2) determine the extent to which ADHD polygenic risk associates with phenotypes beyond ADHD that putatively share its underlying liability and have implications for functional outcome. Methods Participants were 470 youth, ages 7 to 18, consecutively referred for neuropsychiatric evaluation and genotyped on the Illumina Infinium PsychArray Beadchip. Diagnoses made subsequent to study enrollment reflected a range of psychopathology and comorbidity. We determined the burden of ADHD-related common variants in these patients at different significance thresholds from the ADHD PGC-iPSYCH meta-analysis. We then associated this polygenic risk with clinical phenotypes. First, we conducted univariate analyses relating ADHD polygenic risks scores (PRSs) to ADHD diagnoses (none/borderline/full) and symptoms, cognition, and dimensional ratings of aggression and mania. Second, we stratified youth by low, medium, and high ADHD polygenic burden and used a mixed modeling approach to determine whether risk strata associated with distinct multivariate clinical profiles. Third, we determined phenotypically distinct latent classes in these and other youth (N~900) based on eight dimensional measures of psychopathology and examined whether the ADHD-PRS discriminated among them. Results The ADHD-PRS predicted the ADHD diagnosis at seven discovery sample thresholds after correction for potential confounds and multiple testing. The strongest association occurred at discovery sample p Discussion In a multi-diagnostic clinical sample, ADHD polygenic risk showed convergent validity with ADHD phenotypes. Genotype- and phenotype-first analyses also indicated that ADHD polygenic burden associates with traits reflecting cognition and behavioral and emotional dysregulation that extend beyond ADHD and that have implications for functional outcome.