Examining the Etiological Specificity of Genetic Risk for Alcohol Use Disorder and Alcohol Consumption through Symptom-Level Mediation Analyses

Introduction. Polygenic scores (PGSs) for alcohol use disorder (AUD) may capture genetic liability related to alcohol consumption as well as broader transdiagnostic traits, including impulsivity, aggression, and neuroticism. However, the extent to which these traits account for associations between an AUD PGS and the presentation of AUD symptoms remains unclear. Thus, to better understand the genetic influences captured by an AUD PGS, the present study used mediation analyses to examine whether these transdiagnostic traits differentially mediate the relations between PGSs for AUD and individual AUD symptoms while conducting a parallel analysis using a PGS of alcohol consumption to examine differences in the genetic influences underlying AUD and consumption. Method. Participants were drawn from the UCSF Family Alcoholism Study. Separate mediation models examined maximum habitual drinks (N = 1,611) and MMPI-2 Personality Psychopathology Five (PSY-5) scales (N = 1,464) as mediators of the association between AUD/alcohol consumption PGSs and individual AUD symptoms. Covariates included sex, age, and 10 ancestry principal components. Observations were modeled as clustered within families, and a Bonferroni correction was applied (p < .0045). Results. Relations between the AUD PGS and AUD symptoms were fully mediated by maximum habitual drinks and partially mediated by neuroticism for nearly all symptoms, but not by other PSY-5 scales. Relations between the alcohol consumption PGS and AUD symptoms were fully mediated by maximum habitual drinks alone, with no mediation observed through any personality psychopathology measure. Discussion. Current AUD and alcohol consumption PGSs appear to capture genetic liability that operates predominantly through alcohol consumption. However, the AUD PGS may capture genetic influences on broader transdiagnostic factors as well, with neuroticism acting as a partial mediator. Thus, etiological interpretations of the AUD PGS as solely a measure of AUD risk should be made cautiously.