Alcohol consumption during pregnancy and childrens cognitive development: Mendelian randomisation study using the ADH1B genotype

Little is known about the effects of moderate drinking during pregnancy. Randomised trials are not possible and observational studies have problems with confounding and bias. The aim of this study was to investigate the relationship between drinking before recognition of pregnancy early during pregnancy and childrens cognitive develop-ment, using a single nucleotide polymorphism (SNP) of the alcoholdehydrogenase gene ADH1B as a proxy for intrauterine alcohol exposure.This was a Mendelian randomisation study, based on the principle that genotypes are randomly distributed in a population and should not be associated with confounding variables or subject to reporting bias. We used instrumental variable methods to estimate the effect of alcohol consumption before pregnancy recognition on childrens IQ tests at age 8 school results at age 11.We used data from ALSPAC, a longitudinal cohort study of 14,541pregnancies. 7,259 motherchild pairs were genotyped and IQ resultswere available for 4,103 and school test results for 6,432 children.Using observational methods and adjusting for confounding variables,we found that drinking before pregnancy has a positive effect on childrens cognitive outcomes. However, drinking patterns are strongly related tosocial class, education and lifestyle and possibly to unmeasured variables.Using the Mendelian randomisation approach, the relationship between alcohol consumption and childrens outcomes was reversed. The effect ofan increase in frequency of drinking before recognition of pregnancy was an average decrease in IQ scores and school test results at age 11.