Pregnancy and child

Light drinking in pregnancy

“An occasional glass of wine during pregnancy won’t harm a baby’s development,” reported the Daily Mail. The newspaper said, “pregnant women who drink one or two units of alcohol a week may actually find their child is better behaved than if they abstained”.

This large study tested the emotional, behavioural and cognitive development of 11,513 children in the UK when they were five years old. The children’s performance on these tests was compared with the mother’s recollection of her alcohol consumption during pregnancy.

Contrary to the newspaper report, the researchers did not conclude that pregnant women who drink lightly will have better-behaved children. They actually state that their findings indicate that light drinking has no benefits and no harms to children. Also, though this was a large, well-conducted study that used accepted methods, it has several limitations, and the case for no harm from light drinking is not conclusive.

This research does not affect the official UK guidance, which is that alcohol should be avoided during the first three months of pregnancy. Beyond three months, a safe level of alcohol has not been firmly established, though the guidance states there is no evidence of harm from drinking a maximum of 1-2 units once or twice a week.

Where did the story come from?

This study was carried out by researchers from University College London, the University of Essex, Oxford University and the University of Warwick. The research was funded by the Economic and Social Research Council. The study was published in the peer-reviewed Journal of Epidemiology and Community Health.

The Daily Express, Guardian, Daily Mail, BBC News_ and Daily Telegraph_ covered the story. The Express’s suggestion that the research found that “one glass of wine a week while pregnant could help a child’s development in early life” is not justified, and the researchers themselves state they found that light drinking appeared to have no benefits or harms to offspring. This statement may be based on differences in behavioural problems between the groups. However, this difference loses statistical significance once other factors are taken into account.

What kind of research was this?

This cohort study examined whether light drinking during pregnancy affected the risk of social, emotional and cognitive problems in the child at age five.

A cohort study is the best way of assessing whether an exposure (in this case alcohol) is associated with an effect (developmental problems in the child). However, with this type of question, accurately quantifying the amount of alcohol consumed during pregnancy may be difficult, particularly when this is assessed retrospectively, as was the case with this study. Because women who drink during pregnancy and those who do not may also differ in other ways, it is important to take other possible confounding factors (e.g. smoking) into account.

What did the research involve?

This study used data collected in the UK Millennium Cohort Study, a nationally representative study of 11,513 infants born in the UK between September 2000 and January 2002. Households with babies born during this time were identified through the Department of Work and Pensions’ child benefit system. The first survey was conducted when the babies were nine months old. It included questions on the mothers’ drinking during pregnancy, other health-related behaviours, socioeconomic circumstances and household composition.

Participants were grouped according to whether the mother reported:

  • Never drinking (being teetotal)
  • Not drinking in pregnancy
  • Drinking a light amount during pregnancy (1 to 2 units per week or per occasion)
  • Drinking moderately (not more than 3 to 6 units per week or 3 to 5 units per occasion)
  • Being a heavy/binge drinker (7 or more units per week or 6 or more units per occasion).

The researchers were particularly interested in the effects of light drinking in pregnancy.

Later surveys were carried out when the children were two and five years old. At age five, trained interviewers visited the children’s homes and assessed the child’s cognitive development, social and emotional behaviour, socioeconomic factors and the psychosocial environment of the family. Behavioural and emotional assessments were carried out using the Strengths and Difficulties Questionnaire (SDQ), which was completed by the parents.

This is a validated tool assessing five domains of behaviour (conduct problems, hyperactivity, emotional symptoms, peer problems and prosocial behaviour [altruism]). Cognitive assessments were made using age-appropriate tests from the British Ability Scales (BAS), which has subscales assessing vocabulary, picture similarity recognition and pattern construction. To identify children with problems in each of the areas assessed, the researchers used previously defined clinically relevant cut-offs on the SDQ and standardised scores for the BAS subscales.

Many potential confounders were taken into account, including the parents’ lifestyle behaviours, mental health, occupation and parenting style. Only children who were white and were single births (i.e. not twins) were included so that ethnicity and multiple births could not affect results.

What were the basic results?

The researchers found that:

  • 5.9% of mothers were teetotal
  • 60.2% did not drink during pregnancy
  • 25.9% were light drinkers during pregnancy
  • 5.5% were moderate drinkers
  • 2.5% were categorised as heavy/binge drinkers.

When the researchers looked at associations between drinking and other factors, they found that light drinkers were more socioeconomically advantaged compared with mothers in all other categories.

The behavioural and emotional score found that light drinkers were less likely than mothers who did not drink during pregnancy to have children with a high total difficulties score on the SDQ? (6.6% of boys born to light drinkers had a high score compared to 9.6% born to non-drinkers, and 4.3% of girls born to light drinkers had a high score compared to 6.2% born to non-drinkers.

Light drinkers were also less likely to have children with a high hyperactivity score on the SDQ (10.1% of boys born to light drinkers had a high score compared to 13.4% born to non-drinkers. For girls, the figures were 5.5% vs. 7.6%). However, when the calculations were fully adjusted for all potential confounding factors, these differences were no longer statistically significant. No other emotional or behavioural domains on the SDQ showed a statistically significant relationship with light drinking in the mother.

The cognitive assessment found that children born to light drinkers had slightly higher mean (average) cognitive test scores compared with those born to mothers who did not drink during pregnancy. Boys had higher scores in domains of naming vocabulary (a score of 58 in the light drinking group compared to 55 in the group who did not drink in pregnancy), picture similarities (56 versus 55 respectively) and pattern construction (52 versus 50).

Girls born to light drinkers had slightly higher scores for naming vocabulary (58 versus 56) and pattern construction (53 versus 52) compared to girls born to mothers who did not drink during pregnancy. When fully adjusted for confounders, differences remained statistically significant only for boys, in the domains of naming vocabulary and picture similarities.

How did the researchers interpret the results?

The researchers conclude that at age five, children born to mothers who drank 1 to 2 units per week or per occasion during pregnancy were not at increased risk of clinically relevant behavioural difficulties or cognitive deficits compared with children of mothers who did not drink during pregnancy.


This relatively large study collected data on 11,513 UK children who were assessed using validated tests of their cognitive, behavioural and emotional behaviour at age five. The researchers assessed how the children were affected by the mothers’ alcohol consumption during pregnancy (reported when the baby was nine months old). As the researchers concluded, the study does not find any evidence that light drinking in pregnancy poses risks to the behavioural, emotional or cognitive development of the young child. However, there are some important points to take into account:

  • The study carried out many statistical tests on the relationship between five domains of emotional and behavioural development, and three domains of cognitive ability, within the five categories of pregnancy drinking. The tests were done separately for both boys and girls. When the researchers had fully taken into account all the confounders they had identified, the only statistically significant results were a slightly higher score on two domains of cognitive ability (vocabulary and picture similarities) for boys whose mothers drank lightly during pregnancy compared to boys whose mothers did not drink during pregnancy.

Therefore, overall, with only two positive associations found out of a very large number of statistical comparisons, this study provides no clear evidence for the effect of light alcohol consumption, or any other level of alcohol consumption, during pregnancy on the cognitive, emotional and behavioural development of the child. Furthermore, when carrying out such a large number of statistical comparisons, there is a higher likelihood of finding differences by chance. * Another limitation is that the mothers’ alcohol consumption during pregnancy was assessed retrospectively when the baby was nine months old. This may introduce some recall inaccuracy. Some mothers who drank might also feel that reporting their true level of drinking might reflect badly on them. Also, most of the mothers were drinkers normally, but stopped during pregnancy; it may be difficult to say when drinking actually stopped and how this related to the time of conception. * Most of the women in this study did not drink during pregnancy. Those who did, mostly drank small amounts only. As such, only a small number of women were classified as heavy drinkers, and the study may not have reliably detected the effects of heavier drinking during pregnancy. * One of the study’s strengths is that it considered a wide range of potential maternal and child confounders, including medical, lifestyle, socioeconomic, environmental and psychosocial factors. However, there is likely to be an interaction between many of the different factors that can influence a child’s development, and it is hard to pick these apart and remove their effects entirely.

What is the official NICE guidance?

NICE antenatal care guidance (CG62, published June 2010) gives the current recommendations regarding alcohol in pregnancy:

  • Pregnant women and women planning a pregnancy should be advised to avoid drinking alcohol in the first three months of pregnancy because it may be associated with an increased risk of miscarriage.
  • If women choose to drink alcohol during pregnancy they should be advised to drink no more than 1 to 2 UK units once or twice a week (1 unit being equivalent to half a pint of ordinary strength lager or beer, or one shot [25 ml] of spirits. One small [125 ml] glass of wine is equal to 1.5 UK units). Although there is uncertainty regarding a safe level of alcohol consumption in pregnancy, at this low level there is no evidence of harm to the unborn baby.
  • Women should be informed that getting drunk or binge drinking during pregnancy (defined as more than 5 standard drinks or 7.5 UK units on a single occasion) may be harmful to the unborn baby.

NHS Attribution