While heavy drinking is clearly harmful to the unborn baby, often leading to miscarriage, premature birth and foetal alcohol syndrome, the possible effects of light drinking have been less clear.

High quality data on this issue is lacking due to ethical and methodological issues. On the one hand, experiments (clinical trials) in this area are impossible to conduct. Clinical trials would include randomising a group of pregnant women to drinking alcohol, which is clearly unethical. On the other hand, in observational studies we can never be sure whether the results are due to alcohol or other factors, such as wealth or education.

‘One glass is OK, isn’t it?’