‘Researchers from Oxford and Bristol universities looked at the IQ scores of 4,000 children as well as recording the alcohol intake of their mothers,’ reports the BBC. Should I drink or not drink during pregnancy on the basis of this study?
“Even at levels of alcohol consumption which are normally considered to be harmless, we can detect differences in childhood IQ which are dependent on the ability of the fetus to clear this alcohol,” Sarah Lewis, senior lecturer in genetic epidemiology at Bristol University, said in a news release.
The research, published in Plos One, used data from over 4,000 mothers and their offspring, concludes that you shouldn’t. Are they right, the BBC seems to think so.
What did they do?
At 18 weeks’ gestation women were asked to complete a questionnaire on their average amount and frequency of alcohol consumption. At 32 weeks of gestation women completed another questionnaire. Cognitive tests were carried out at 8 year of age using a shortened version of the Wechsler Intelligence Scale for Children (WISC-III)
Ten Single Nucleotide Polymorphisms genes were selected because they had been shown to be associated with alcohol and related genes. Associations were tested between WISC score at age 8 and genotype using linear regression models. These models were adjusted for a number of factors.
What did they find
The mean age-adjusted WISC score among the 6,196 eligible children who completed the test at age 8 was 105. From 20 different alleles they reported five gene variants, involved in alcohol metabolism, were associated with an 8 year olds cognitive ability. The associations between child’s genotype and outcome were only present among those reporting drinking alcohol in moderation during pregnancy and not those who didn’t drink at all.
The research suggested there was a dose response for those drinking during pregnancy but the drinking group had a much higher baseline IQ (107.5 vs. 103.1). Overall, the problem with this paper is the sheer number of statistical tests being undertaken, which generally leads to spurious results. It is misleading, and not good practice, to just look at reported statistical significance, irrespective of how small the p value is and then infer causality.
What follows is why.
The WISC III for children scale is an individually administered and standardised test. A score of 90-109 is classified as average. Scores between 120-129 are classed as “very high,” 110-119 “bright normal,” 85-89 low average, and 70-84 low average. Although the differences reported between the groups in the paper are statistically significant they are clearly not significant in terms of the classification system. Not even close.
In addition, the regression model used could, and should, have been adjusted for all the factors not just individually.
This work has all the hallmarks of data mining, no prescified hypothesis, too many statistical tests, small insignificant differences found and inappropriate statistical methods employed. Therefore if you are asked for advice on this topic, which as a GP I often am, the current study adds little to current understanding.
The advice, and the answer to our question is, drink moderately in pregnancy.