Alcohol in pregnancy: what the evidence on harm actually shows

There is no established safe amount of alcohol in pregnancy, and health authorities in both the UK and the US advise women who are pregnant or trying to conceive not to drink at all [s1][s2]. That guidance is precautionary rather than a claim that a single drink is proven to harm: heavy and binge drinking clearly cause fetal alcohol spectrum disorder, but the evidence on light drinking is genuinely thin, and what exists cannot rule out harm [s3][s4][s5].

What fetal alcohol harm is

Alcohol crosses the placenta, and prenatal exposure can damage the developing brain and body. The umbrella term fetal alcohol spectrum disorder (FASD) covers a range of lifelong effects on growth, facial development, cognition and behaviour; fetal alcohol syndrome (FAS) is the most severe, fully expressed form [s4][s5]. Alcohol use in pregnancy is the direct cause of FAS, and the condition is, in principle, entirely preventable [s5].

It is more common than many assume. A meta-analysis of 24 studies covering 1,416 diagnosed children estimated the global prevalence of FASD among children and youth at 7.7 per 1,000 — roughly 1 in 130 — but with enormous variation [s4]. The WHO European Region had the highest regional prevalence at 19.8 per 1,000, and the WHO Eastern Mediterranean Region the lowest at 0.1 per 1,000 [s4]. Among individual countries, South Africa topped the list at 111.1 per 1,000, followed by Croatia at 53.3 and Ireland at 47.5 [s4]. That variation tracks how much, and how commonly, women drink.

A companion analysis quantified the drinking itself: globally, an estimated 9.8 percent of women drink alcohol during pregnancy (95% confidence interval 8.9 to 11.1), and the prevalence of FAS was put at 14.6 per 10,000 people (95% CI 9.4 to 23.3) [s5]. Linking the two, the authors estimated that about 1 in every 67 women who drink in pregnancy delivers a child with FAS, translating to roughly 119,000 children born with FAS worldwide each year [s5].

The hard part: what about light drinking?

The scientifically honest answer is that the evidence on low-level drinking is remarkably sparse. A systematic review pooling 24 cohort and two quasi-experimental studies looked specifically at light drinking — defined as up to 32 grams of alcohol a week, about four small drinks — compared with abstaining [s3]. For most outcomes there were too few data to combine or draw firm conclusions [s3]. Where meta-analysis was possible, babies of mothers who drank up to 32 g/week had higher odds of being small for gestational age (summary odds ratio 1.08, 95% CI 1.02 to 1.14) and of preterm birth, though the preterm estimate was also compatible with no effect (OR 1.10, 95% CI 0.95 to 1.28) [s3].

The authors' conclusion is the crux of the whole debate: because evidence of the effects of drinking at this level is so limited, guidance "could advise abstention as a precautionary principle but should explain the paucity of evidence" [s3]. In plain terms, the absence of proof of harm from light drinking is not the same as proof of safety — and the studies that would settle it largely do not exist, partly because they are hard and unethical to run.

What the guidelines say

Public-health bodies have resolved that uncertainty on the side of caution. The UK Chief Medical Officers' guidance states that if a woman is pregnant or could become pregnant, the safest approach is not to drink alcohol at all, to keep risk to the baby to a minimum [s1]. The American College of Obstetricians and Gynecologists is similarly categorical, advising that there is no known safe amount, no safe time, and no safe type of alcohol to drink during pregnancy [s2]. Both frame abstinence as the way to eliminate a preventable risk rather than as a verdict that any exposure is proven to cause harm [s1][s2].

Why guidelines and headlines so often seem to disagree on alcohol is itself a recurring theme; we examine it in why national drinking guidelines can't agree on a safe level. The wider evidence on alcohol and disease is covered in what the evidence says about alcohol and cancer and what happened to the alcohol "J-curve". For another exposure where pregnancy advice hinges on imperfect data, see our review of paracetamol in pregnancy.

How to read this

The evidence supports a clear, two-part message. Heavy and binge drinking in pregnancy demonstrably harm the developing child and drive a preventable, globally common disability [s4][s5]. For light drinking, the data are too sparse to declare either harm or safety, which is exactly why authorities advise abstaining as the cautious default [s3][s1][s2]. This article is informational and not medical advice; questions about alcohol before or during a pregnancy should be discussed with a qualified clinician or midwife.