Identifying suitable instrument for Mendelian randomization studies of the consequences of prenatal alcohol exposure in European populations remains a challenge. The most natural choice of candidates is alcohol dehydrogenase (ADH) genes, which catalyze the first step of alcohol metabolism oxidizing ethanol into acetaldehyde. A role for these variants in alcohol intake, metabolism and dependence has been suggested by both genome-wide linkage (18–21) and candidate gene (22–25) studies, and in particular a non-synonymous variant in ADH1B has been implicated. The rare variant is thought to encode an enzyme which accelerates ethanol clearance (26) (based on studies in vitro), but there is limited evidence for association with alcohol metabolism in vivo (22,27,28,29). The extent to which this SNP is associated with alcohol intake among pregnant women is unclear. On the one hand, the pressures to cut down or quit drinking alcohol (30–32) might increase the relative contribution of genetic variation to alcohol intake and, on the other; pregnancy might result in metabolic changes and modify the effect of the ADH1B and other ADH variants. A systematic review on the incidence of fetal
