New evidence supports advice not to drink alcohol in...

"Drinking alcohol during pregnancy really does lead to poor brain function in babies, study confirms," reports the Mail Online.

The Chief Medical Officers for the UK currently advise women not to drink any alcohol at any point during pregnancy, to keep risks to the baby to a minimum. A new review of studies found some evidence to suggest that drinking alcohol in pregnancy may affect babies' thinking abilities in later life, and possibly cause lower birthweight.

However, the researchers said the studies included in their review could have biased results, and the results should be treated with caution. The review does not answer the question of whether there is a low level of alcohol consumption that is safe during pregnancy.

Official UK advice about alcohol in pregnancy remains the same. The evidence from this study adds to concerns that alcohol may harm babies, so the safest option is still to not drink alcohol at all at any stage of the pregnancy.

Find out more about alcohol in pregnancy.

Where did the story come from?

The researchers who carried out the review were from the University of Bristol. The study was funded by the National Institute for Health Research and the Medical Research Council.

It was published in the peer-reviewed International Journal of Epidemiology on an open access basis so is free to read online.

The headline used by the Mail Online overstates the certainty of the results of the review. Their report says that the study showed "drinking alcohol at any point in pregnancy leads to poorer brain function in babies". However, what the study showed was that there was some evidence of a higher risk of lower brain function for babies exposed to alcohol.

The Daily Telegraph report makes it clearer that there is uncertainty over the results.

What kind of research was this?

This was a systematic review of quasi-experimental studies of alcohol and pregnancy. Quasi-experimental means the researchers tried to find ways to get around the problems of observational studies.

Randomised controlled trials are the best way of looking at the direct effect of an exposure. For obvious ethical reasons, women cannot be randomised to drink alcohol during pregnancy. Therefore, the research usually has to be observational.

In observational studies, researchers compare what happens to babies born to children whose mothers report drinking alcohol at varying levels. The problem with this is that lots of things – such as education, income, age, tobacco or drug use and general health – may vary according to whether a woman drinks alcohol in pregnancy.

These confounding factors may also affect the health of her child, so it is hard to untangle the direct effect of alcohol. Observational studies have had mixed results when looking at low to moderate alcohol drinking in pregnancy.

Quasi-experimental studies try to find ways to reduce some of this confounding from other influences. They may include:

studies of the same woman in different pregnancies, where her alcohol consumption varied
studies grouping women according to certain genetic variations, which act as a marker for whether they drink alcohol
“natural experiment" studies that looked at things outside a woman's control, such as changes in legal drinking age or alcohol pricing

What did the research involve?

The researchers looked for any randomised controlled trials or quasi-experimental studies that assessed the effects of alcohol in pregnancy on outcomes during pregnancy, or after the child's birth.

They identified a total 23 studies, which grouped together by type included:

1 randomised controlled trial – which looked at the effects of helping women who drank alcohol to stop during pregnancy
9 studies looking at women carrying different genes that affect their ability to make the enzyme alcohol dehydrogenase (ADH) – which breaks down alcohol
7 natural experiment studies – which mainly looked at the effect of changes to minimum age of alcohol consumption
6 "negative control" studies – for example if you compared alcohol consumption in mothers and fathers, the father would be the "negative control" as you would not expect his alcohol consumption to have any effect on the baby

The researchers intended to pool results from studies of the same type that looked at the same outcome, but this was only possible for 2 studies. Instead, they described the results of the groups of studies and looked to see whether the results were consistent.

The researchers looked at potential risk of bias for all the studies, to see how reliable the results are likely to be.

What were the basic results?

The researchers found some evidence from the 23 studies that alcohol may affect babies' thinking ability and childhood achievement. They found weaker evidence that they may affect birthweight.

However, some of the studies found no evidence of these outcomes.

Out of the 9 genetic variation studies:

2 studies found some evidence of a higher risk of foetal alcohol syndrome among children more likely to have been exposed to alcohol during pregnancy
1 study found evidence that children's IQ was lower at age 8 if they were more likely to have been exposed to alcohol
2 studies found children did better at school if exposed to less alcohol in pregnancy, but there was no effect on IQ

Out of the 2 pregnancy comparison studies, 1 found more conduct problems in children exposed to alcohol. Out of the 4 parental comparison studies, 1 found more evidence that mother's, rather than father's, drinking was linked to poorer school results at age 11.

Out of the 7 natural experiments:

3 US studies found that children born after the minimum alcohol drinking age was lowered (so people could drink at a younger age) were more likely to have low birthweight or be born early
1 study found that higher alcohol taxes were linked to increased birthweights and responsiveness of babies at birth

In the randomised controlled trial of women given help to not drink during pregnancy, there was no evidence that women given such help decreased their drinking.

The researchers said that all of the studies carried some potential risk of bias, meaning the results should be treated with caution.

How did the researchers interpret the results?

The researchers said their research had found "a modest body of evidence suggestive of a detrimental effect on cognitive outcomes and infant birthweight".

They said this "will strengthen the evidence base for the abstinence recommendation" – which is not to drink at all in pregnancy.

Conclusion

Not much has changed as a result of this study. The recommendation to pregnant women remains to avoid alcohol completely throughout pregnancy. The study adds some additional weight to the evidence that led to that recommendation being made by the Department of Health in 2016.

The review highlights how difficult it is to judge the impact of just one factor, such as alcohol consumption, on the outcome of a pregnancy.

It is not possible to carry out a trial in which some women are randomly assigned to drink alcohol while others are assigned not to drink, because that would be unethical. So this review looked for special types of observational studies that have used other methods to try and reduce the influence from other environmental, health or lifestyle factors. But all of these studies rely on assumptions, which means we have to be cautious about the results.

For example, if a mother drinks during one pregnancy but not another, we may assume that everything else about the mother's behaviour during pregnancy was the same – but we cannot know this for sure.

Many of the studies included were quite small, and the variation in methods and outcomes measured meant that few of the studies could be combined.

Despite these limitations, the overall weight of the review supports the understanding that alcohol in pregnancy may increase the chance of harm to babies, although we do not know how much alcohol poses a risk. For this reason, the advice about alcohol and pregnancy remains the same.

Read more about drinking alcohol during pregnancy.

Page last reviewed: Tue Feb 2020 Next review due: Tue Feb 2020