Second, we have not accounted for parental death, which can be a risk factor for AUD (11). Furthermore, in the absence of two parents, having a single affected parent may differentially increase risk of young adult AUD. Third, parental EB prior to a child’s birth, which could be considered informative of genetic risk that might not manifest after the child’s birth, could not be assessed for all cohorts, and therefore was not included as a covariate. However, results were unchanged when this variable was accounted for in analyses replicated in only the youngest parental cohort. Fourth, we did not incorporate census data, so these analyses do not account for the fact that not all children lived with both parents during all time frames included. However, we conducted post hoc analyses restricted to families where census data confirmed that the child lived with both parents during a given time frame, and results did not substantively differ. Fifth, data availability precluded an examination of later adult AUD.
